Rhetorical Tropes in Political Discourse: Relevant Website

Rhetorical Tropes in Political Discourse 597
Mann W C, Matthiessen C M I M & Thompson S A (1992).

Rhetorical structure theory and text analysis. In Mann
W C & Thompson S A (eds.) Text description: diverse
analyses of a fund raising text. Amsterdam: John Benjamins. 3978.
Mann W C & Moore J A (1981). Computer generation of
multiparagraph text. American Journal of Computational Linguistics 7(1), 1729.
Mann W C & Thompson S A (1986). Relational propositions in discourse. Discourse Processes 9(1), 5790.
Mann W C & Thompson S A (1988). Rhetorical structure
theory: toward a functional theory of text organization.
Text 8(3), 243281.
Marcu D (2000a). The rhetorical parsing of unrestricted
texts: a surface-based approach. Computational Linguistics 26(3), 395448.
Marcu D (2000b). The theory and practice of discourse
parsing and summarization. Boston: The MIT Press.
Marcu D, Carlson L & Watanabe M (2000). The automatic translation of discourse structures. Proceedings of the
Annual Meeting of the North American Chapter of the
Association for Computational Linguistics (NAACL).
http://www.isi.edu/marcu/papers/dmt-naacl2000.ps.
Martin J R (1992). English text: systems and structure.
Amsterdam: John Benjamins.
Matthiessen C M & Thompson S A (1989). The structure
of discourse and subordination. In Haiman J &
Thompson S A (eds.) Clause combining in grammar and
discourse. Amsterdam: John Benjamins.
Moore J D & Paris C L (1988). Constructing coherent
texts using rhetorical relations. Proceedings of the
Tenth Annual Conference of the Cognitive Science Society. Montreal: Cognitive Science Society. 637643.
Moore J D & Pollack M E (1992). A problem for RST: the
need for multi-level discourse analysis. Computational
Linguistics 18(4), 537544.
ODonnell M (1997). Variable-length on-line document
generation. Proceedings of the Flexible Hypertext
Workshop of the Eighth ACM International Hypertext
Conference. UK: Southampton. 7882.
Ro sner D & Stede M (1992). Customizing RST for the
automatic production of technical manuals. In Dale R,
Hovy E, Ro sner D & Stock O (eds.) Aspects of automated natural language generation. Proceedings of the
6th International Workshop on Natural Language Generation. Trento, Italy, April 1992. Berlin: Springer.
199214.
Sacerdoti E (1973). Planning in a hierarchy of abstraction
spaces. Proceedings of IJCAI III: International Joint
Conference on Artificial Intelligence. 412422.
Sanders T J M, Spooren W P M & Noordman L G M
(1992). Towards a taxonomy of coherence relations.
Discourse Processes 15(1), 136.
Taboada M T (2004). Building coherence and cohesion:
task-oriented dialogue in English and Spanish. Amsterdam: John Benjamins.
Taboada M & Mann W C (2005). Rhetorical structure
theory: looking back and moving ahead. Discourse Studies (to appear).
Taboada M & Mann W C (2006). Applications of rhetorical structure theory. Discourse Studies (8).
Vander Linden K (2000). Natural language generation. In
Jurafsky D & Martin J (eds.) Speech and language
processing: an introduction to speech recognition,
computational linguistics and natural language processing. New Jersey: Prentice-Hall. 763798.
Webber B, Knott A, Stone M & Joshi A (1999). Discourse
relations: a structural and presuppositional account using
lexicalized TAG. ACL 99: Proceedings of the 37th
Annual Meeting of the American Association for
Computational Linguistics. American Association for
Computational Linguistics, University of Maryland.
4148.
Relevant Website
http://www.sfu.ca/rst/ The RST website includes definitions of all the RST relations, many example analyses,
and translations of basic introductions in French and
Spanish.
Rhetorical Tropes in Political Discourse

M Reisigl, University of Vienna, Vienna, Austria
2006 Elsevier Ltd. All rights reserved.
Classical Rhetorical and Actual

Characterization of Tropes
The Greek substantive tropo s, originally meaning
turn or direction, is derived from the verb tre pein,
which signifies to turn (over/round). The classical
rhetorical term encompasses all figures of speech

that involve a turn of meaning (i.e., a linguistic
transference from one conceptual sphere to another).
Conventionally, tropes are considered to be conventionalized means of expression of so-called improper
speech or nonliteral speech, which is characterized
by a (poetically) licensed difference between the ordinary literal and the extraordinary intended meaning of a speech. This tropological theory of deviation

Mann W C, Matthiessen C M I M & Thompson S A (1992).
Rhetorical structure theory and text analysis. In Mann
W C & Thompson S A (eds.) Text description: diverse
analyses of a fund raising text. Amsterdam: John Benjamins. 3978.
Mann W C & Moore J A (1981). Computer generation of
multiparagraph text. American Journal of Computational Linguistics 7(1), 1729.
Mann W C & Thompson S A (1986). Relational propositions in discourse. Discourse Processes 9(1), 5790.
Mann W C & Thompson S A (1988). Rhetorical structure
theory: toward a functional theory of text organization.
Text 8(3), 243281.
Marcu D (2000a). The rhetorical parsing of unrestricted
texts: a surface-based approach. Computational Linguistics 26(3), 395448.
Marcu D (2000b). The theory and practice of discourse
parsing and summarization. Boston: The MIT Press.
Marcu D, Carlson L & Watanabe M (2000). The automatic translation of discourse structures. Proceedings of the
Annual Meeting of the North American Chapter of the
Association for Computational Linguistics (NAACL).
http://www.isi.edu/marcu/papers/dmt-naacl2000.ps.
Martin J R (1992). English text: systems and structure.
Matthiessen C M & Thompson S A (1989). The structure
of discourse and subordination. In Haiman J &
Thompson S A (eds.) Clause combining in grammar and
discourse. Amsterdam: John Benjamins.
Moore J D & Paris C L (1988). Constructing coherent
texts using rhetorical relations. Proceedings of the
Tenth Annual Conference of the Cognitive Science Society. Montreal: Cognitive Science Society. 637643.
Moore J D & Pollack M E (1992). A problem for RST: the
need for multi-level discourse analysis. Computational
Linguistics 18(4), 537544.
ODonnell M (1997). Variable-length on-line document
generation. Proceedings of the Flexible Hypertext
Workshop of the Eighth ACM International Hypertext
Conference. UK: Southampton. 7882.
Rosner D & Stede M (1992). Customizing RST for the
automatic production of technical manuals. In Dale R,
Hovy E, Rosner D & Stock O (eds.) Aspects of automated natural language generation. Proceedings of the
6th International Workshop on Natural Language Generation. Trento, Italy, April 1992. Berlin: Springer.
199214.
Sacerdoti E (1973). Planning in a hierarchy of abstraction
spaces. Proceedings of IJCAI III: International Joint
Conference on Artificial Intelligence. 412422.
Sanders T J M, Spooren W P M & Noordman L G M
(1992). Towards a taxonomy of coherence relations.
Discourse Processes 15(1), 136.
Taboada M T (2004). Building coherence and cohesion:
task-oriented dialogue in English and Spanish. Amsterdam: John Benjamins.
Taboada M & Mann W C (2005). Rhetorical structure
theory: looking back and moving ahead. Discourse Studies (to appear).
Taboada M & Mann W C (2006). Applications of rhetorical structure theory. Discourse Studies (8).
Vander Linden K (2000). Natural language generation. In
Jurafsky D & Martin J (eds.) Speech and language
processing: an introduction to speech recognition,
computational linguistics and natural language processing. New Jersey: Prentice-Hall. 763798.
Webber B, Knott A, Stone M & Joshi A (1999). Discourse
relations: a structural and presuppositional account using
lexicalized TAG. ACL 99: Proceedings of the 37th
Annual Meeting of the American Association for
Computational Linguistics. American Association for
Computational Linguistics, University of Maryland.
4148.
Relevant Website
http://www.sfu.ca/rst/ The RST website includes definitions of all the RST relations, many example analyses,
and translations of basic introductions in French and
Spanish.
Rhetorical Tropes in Political Discourse

M Reisigl, University of Vienna, Vienna, Austria
Classical Rhetorical and Actual

Characterization of Tropes
The Greek substantive tropos, originally meaning
turn or direction, is derived from the verb trepein,
which signifies to turn (over/round). The classical
rhetorical term encompasses all figures of speech

that involve a turn of meaning (i.e., a linguistic
transference from one conceptual sphere to another).
Conventionally, tropes are considered to be conventionalized means of expression of so-called improper
speech or nonliteral speech, which is characterized
by a (poetically) licensed difference between the ordinary literal and the extraordinary intended meaning of a speech. This tropological theory of deviation
598 Rhetorical Tropes in Political Discourse
or substitution is nowadays often criticized for mistakenly assuming that tropes are deviations from linguistic normality and can thus simply be replaced by
proper expressions or phrases. However, tropes are,
in fact, very common in everyday language, and
there is no such thing as an original verbum proprium
or substituendum that perfectly corresponds to the
trope (cf., among others, Kienpointner, 1999: 6668).
In contrast to other figures of speech (e.g., figures
of repetition like anaphora, epiphera, alliteration,
assonance, consonance, and parallelism), tropes are
said to be more closely related to content than to form
or structure. Concerning the relationship between
what is said or written and what is meant, Heinrich
Lausberg (1990a: 6479) distinguishes between two
types of tropes.
Tropes of shifting boundaries (Grenzverschiebungstropen) consist of moving the borders of neighboring
semantic fields or the borders within one and the
same semantic field. These tropes are determined by
a relationship of inclusion or (factual) contiguity between what is said or written and what is meant. They
are, accordingly, divided into two subcategories. Periphrasis, litotes, hyperbole, emphasis, antonomasia,
and synecdoche are assigned to the first subcategory.
They are constituted by shifting the boundary within
a semantic sphere. The second subcategory involves a
relationship of adjacency. It is prototypically represented by metonymy.
The second type of tropes is labeled leaping tropes
(Sprungtropen). In comparison to tropes that involve
shifting boundaries, leaping tropes are grounded on a
leap from one semantic sphere to another sphere
that is not adjacent to the first one. Traditionally,
metaphor, allegory, and irony are taken as leaping
tropes. Their intended meaning is to be found in a
semantic domain that clearly differs from what is actually said or written. In concrete rhetorical analyses,
a neat distinction between the two types of tropes as
well as between the single tropes allocated to these
two types is not always feasible, as simultaneous
tropic membership of specific linguistic realizations is
not unusual (see Fontanier, 1977; Morier, 1989;
Lausberg, 1990a, 1990b; Groddeck, 1995; Plett,
2001, for various proposals to classify tropes; see
Goossens, 1995, for multiple tropic membership,
discussed with the example of metaphtonymy,
the simultaneous combination of metaphor and
metonymy).
Antique rhetoric ascribes primarily an ornamental
function to tropes. It considers them to be deviations
from ordinary language for the sake of adornment.
This view of tropes as purely decorative linguistic
appendices that are added to the linguistic raw
material in the rhetorical production stage of
elocution to give glamour to ones speech has widely

been overcome by a reassessment that sees language
as basically tropic. Nowadays, it is almost commonplace in modern rhetoric, linguistics, language
philosophy, and cognitive science that all areas of
language are pervaded with tropes, and that tropes,
like metaphors, metonymies, and synecdoches, function as elementary cognitive principles that often
unconsciously shape and structure human perception and thinking. Tropes are no longer exclusively
associated with the stage of elocution; rather, they are
also related to the stage of inventio and, of course, to
processes of the speechs reception or apperception.
Political Discourses
The contention of the tropes pervasiveness in language holds also for the realm of political language,
which generally speaking evolves in a tension
between the preservation and transformation of
power relations, public decision making, and problem solving, as well as political order. Many names of
political institutions and collective political actors
have their origin in metaphorization and metonymization (e.g., government, parliament, minister,
and ministry). Political discourses are full of tropes
related to the three political dimensions of form (i.e.,
polity, the basic order related to political norms,
institutions, system, and culture), content (i.e., policy,
the policy-field-related planning and acting concerning
the identification of political problems, the development and implementation of political programs, and
the evaluation and correction of their implementation), and process (i.e., politics, the political competition related to the conflicts between political actors
and to the achievement of followers and power). In all
fields of political action within and across which
discourses manifest themselves politicians employ
these figures of speech as effective rhetorical means of
constructing, representing, and transforming political
reality, as well as a means of political persuasion.
Fields of action (cf. Girnth, 1996) may be understood as places of social forms of practice (Bourdieu,
1991: 74) or, in other words, as frameworks of social
interaction (Reisigl, 2003: 148). The spatiometaphorical distinction among different fields of action can
be understood as a differentiation between various
functions or socially institutionalized aims of discursive practices. Among the political fields are the lawmaking procedure; the formation of public attitudes,
opinions, and will; the party-internal formation of
attitudes, opinions, and will; the interparty formation
of attitudes, opinions, and will; the organization of
international and (especially) interstate relations;
political advertising; the political executive and
administration; and the various forms of

political control (for more details, see Reisigl, 2003:
128142). A political discourse about a specific topic
may have its starting point within one field of action
and proceed onward through another one. In this
respect, discourses and discourse topics spread to
different fields and discourses. They cross between
fields, overlap, refer to each other, or are in some
other way sociofunctionally linked with each other
(Reisigl & Wodak, 2001: 36 f.).
In the given context, discourse is to be seen as
topic-related concept that, among others, involves
argumentation about validity claims such as truth
and normative validity, as well as pluri-perspectivity
constituted by the discursive participation of various
social actors who assume different points of view
(Reisigl, 2003: 92; for a mono-perspectivist conceptualization of discourse, see Fairclough, 1995: 14).
Discourses represent semiotic social practices that are
both socially constitutive and socially constituted.
A political discourse about a particular topic can be
understood as a complex bundle of simultaneous and
sequential interrelated linguistic acts. These manifest
themselves within, and across, the above-mentioned
fields of political action as thematically interrelated
and problem-centered semiotic (e.g., oral or written)
tokens, very often as texts that belong to specific
semiotic types (i.e., textual types or genres), which
serve particular political purposes (see Reisigl &
Wodak, 2001: 36).
Rhetorical tropes fulfill many different purposes
in political discourses, especially in regard to positive political self-presentation and negative political
other-presentation. With respect to the ideational metafunction of language, these tropes help to invent or
construct a political reality (e.g., via incorporating
metaphors or allegories like the boat or ship metaphor/allegory, referring to a state), to reduce complexities by simplistic categorization and imaginary
representation of political reality (e.g., via color
metaphors referring to alleged human races), to
vivify, personify, and illustrate abstract or unclear
political ideas (e.g., via personifications and other
animalizing metaphors), to selectively foreground
specific traits of political entities or reality sectors
(e.g., via particularizing synecdoches), and to background or hide specific political aspects, actors, or
actions (e.g., via metonymies that represent persons
by place names).
With respect to the interpersonal metafunction of
language, these tropes are employed to promote the
identification with single or collective political
actors (e.g., leaders, parties, states, or nations), as
well as with their political aims and ideologies
(e.g., via the heart metaphor relating to a specific
nation-state within a confederation of states); to promote in-group solidarity (e.g., via family or kinship
metaphors, such as the androcentric metaphor of
brotherhood), referring to the imagined community
of a nation or of politically united nations; to support
out-group segregation and discrimination (e.g., via
deprecating animalizing metaphors like parasite,
rat, and vermin); to create a feeling of security by
suggesting stability and order (e.g., via construction
metaphors like house, referring to a state); to create
a feeling of insecurity by suggesting chaos, disorder,
danger, and threat (e.g., via flood or wave metaphors
referring to immigrants); to justify, legitimize, or delegitimize specific political actions or their omission
(e.g., via metaphors of gain or cost relating to the
consequences of a specific action or omission of action); and to mobilize political followers to perform
specific actions which can culminate in physical or
military aggression, such as war (e.g., via instigating
militarizing metaphors or rape and monster metaphors relating to the enemy). Some of these functions will be elaborated on in greater detail in the
following section.
Toward a Tropology of the Political

For a long time, analysis of political language has
merely been carried out as a side job both on the part
of political scientists interested in issues of language
and on the part of linguists interested in political
issues. As a consequence, the respective studies in
linguistics and political science have frequently been
amateurish with respect to theory, methods, and
methodology. This critique also relates to the study
of tropes in political discourses. Linguistic, and
especially rhetorical, approaches often have suffered
from politological ignorance and did not venture
beyond a positivist identification and enumeration
of figures of speech, as well as a very general characterization of the persuasive or manipulative potential of single tropes, especially of metaphors. The
analyses of tropes in political science did, however,
often not exceed everyday linguistic triviality when
selectively looking at linguistic images with political
significance.
A remedy for the many shortcomings in the area
can be found in a transdisciplinarian politolinguistic
approach that tries to connect and synthesize rhetoric, linguistic discourse analysis, and political science.
Such an approach should theoretically rely on actual
concepts in political science, as well as on rhetorical
and discourse analytical categories such as those
outlined in the Classical Rhetorical and Actual Characterization of Tropes and Political Discourses
sections. With respect to the topic in question,
politolinguistics should build a tropology of the political (i.e., a theory of tropes that systematizes and
explains the political and linguistic functions or purposes of tropic language in polity, policy, and politics;
for a first sketch, see Reisigl, 2002, and Reisigl, 2003:
237258). In what follows, this tropology is outlined
with the example of three of the four so-called
master tropes (see Burke, 1969: 503517); that is
to say, with the example of metaphor (including
personification and allegory, the latter being conceived as continuous, expanded metaphor), metonymy, and synecdoche (including antonomasia as a
special form of synecdoche). The fourth alleged
master trope, irony, will be dropped, as its character
is very different from that of the three basic tropes. In
contrast to metaphor, metonymy, and synecdoche,
irony is much more heterogeneous. It often involves
the prosodic, gestural, or facial-expressive dimension
of language, which enables one to recognize that
what has been said is not to be taken literally, but
ironically. To explicate the complex character of
irony would go beyond the scope of this introductory
article.
In addition, in the present context, for reasons of
limited space, no attention can be paid to other linguistic occurrences traditionally also apostrophized
as tropes. Periphrasis, litotes, hyperbole, and emphasis are as is irony much more heterogeneous
than metaphor, metonymy, and synecdoche. As rhetorical phenomena, they do often derive from the
three basic tropes or manifest themselves in manifold
tropic ways. Periphrasis, for instance, can linguistically be realized as antonomasia (which is characterized as a descriptive phrase standing for an individual
to be identified with a proper name or, conversely, as
a proper name standing for a general character trait),
synecdoche, etymology (which is based on the principle of nomen est omen), metaphor, allegory, metonymy, euphemism, descriptio and definitio, and so
forth (Plett, 2001: 91). However, some of these
other tropes frequently involve linguistic dimensions
that are not genuinely related to tropes, such as the
so-called trope of emphasis, for example, which
comprises numerous phonological, morphosemantic,
and syntactic characteristics that are not to be seen as
tropic qualities (Reisigl, 1999: 197199). To take into
consideration all these various aspects and their complex interplay would be beyond the scope of this
overview.
Metaphors
Of all tropes, metaphors certainly receive the most

attention in the analysis of political discourses. Many
studies have been carried out in the last decades
that focus on metaphorical speech in different

political and historical contexts (cf., among many
others, Lakoff, 1991, 1992; Chilton & Ilyin, 1993;
Chilton & Lakoff, 1995; Scha ffner, 1995, 2002;
Chilton, 1996; Semino & Masci, 1996; Bo ke, 1997,
2002; Musolff, 2000; El Refaie, 2001; Klein, 2002;
Panagl & Stu rmer, 2002; Stu rmer, 2002).
Theoretically, metaphors can be understood as impertinent predications (Ricur, 1986), in the sense
that an expression is semantically incompatible with
the context of meaning in which the expression is
uttered. Metaphor establishes a similarity between
two different semantic domains. As it is a device for
seeing something in terms of something else (Burke,
1969: 503), it is most closely connected with the
question of perspectivation.
Many different types of metaphor occur in political
discourses. Some metaphors transform inanimate
objects into animate ones, whereas others transform
animate objects into inanimate ones. Some metaphors predicate onto animate objects the quality of
other animate objects, whereas others predicate
onto inanimate objects the quality of other inanimate objects. Some transform abstract entities into
concrete ones that are sensually perceivable, and
others merge different human senses synesthetically
or into catachresis (cf. Plett, 2001: 101104). Many
metaphors in political discourses transfer aspects
from the economic to the political domain, and
many from the domain of private life and family life
onto public life and collective actors. Various metaphors project elements from the domains of sports
onto the political domain. All these and many other
metaphorical types (including those that will be
mentioned in the following explications, which for
the most part are also to be found in Reisigl &
Wodak [2001: 5860]), should be taken into consideration by an elaborated phenomenology of metaphors
in political discourses.
Personifications or anthropomorphizations are metaphors that bring together and link two different semantic fields, one with the semantic feature (human),
the other bearing the semantic feature (human)
(cf. Lakoff & Johnson, 1980: 34). Personifications
are rhetorically used to give a human form, to humanize inanimate objects, abstract entities, phenomena,
and ideas. They play an important role in animating
imagined collective subjects as, for example, races,
nations, ethnicities, states, state unions,
and many other political collective actors (cf.
Goodwin [1990] for the question of engendered
nations). Their apparent concreteness and vividness
often invites hearers or readers to identify or to feel
solidarity with or against the personified political
entity.
In addition to personifications, there are many

other forms of metaphors that are important in
referentially and predicationally constructing ingroups and out-groups in politics and polity,
whether they are imagined as races, nations,
ethnicities, states, and tribes or as specific racialized, national, ethnic, or religious majorities or
minorities. Many of these metaphors function as
collective symbols. Very often, they are simultaneously used both as metaphors and representative
synecdoches (cf. Gerhard & Link, 1991: 18).
In discourses about races, nations, and ethnicities, these metaphors and synecdoches, but also the
respective metonymies, are almost always connected
with specific dychotomic, oppositional predications
that can form networks of semantic isotopes that
help politicians to polarize and divide the world of
social actors into black and white and into good
and bad. The most frequent and conspicuous of these
predications to real or imagined social actors and
political systems that serve positive self-presentation
and negative other-presentation are those of singularity/uniqueness/distinctiveness or individuality, of
identity or similarity, of collectivity, of difference, of
autonomy/independence, of heteronomy/dependency,
of continuity, of discontinuity, of (social) inclusion,
of integration, of union or unity, of (social) cohesion, of (social) exclusion, of fragmentarization, of
multiplicity, and of dissolution.
Apart from these rather abstract identity-related
attributions realized by metaphors, national, racial,
and ethnic stereotypes or characters are predicated
on the basis of metaphors, relying on the collective
symbolic concepts of materiality and body; of material status (e.g., the thermostatic status of warm
versus cold) and states of matter (solid, fluid, and
gaseous); of material qualities like weight (heavy
versus light); of spatiality, spatiodynamics, and temporality (fast, fast-moving, ephemeral, persistent, tenacious, lively, mobile, flexible, slowly,
inert, and lethargic); and of the five sensorial concepts of visuality (fair, pale, clear, and transparent versus dark, gloomy obscure), audibility
(harmonious versus loud and noisy), tactile sensation (hard versus soft), olfactority (nice-smelling
versus stinking), and taste (tasteful versus
tasteless).
Racializing, nationalizing, and ethnicizing metaphors of spatiality are primarily ordered around
the symbolically and evaluatively loaded binary
oppositions of internal versus external or internality versus externality, of height/top and up
versus bottom and down/low, of foundations/
profundity/ ground versus bottomlessness, of
superficiality and flatness versus depth, of center
versus periphery/margin, and of boundary, limit,

and extension/expansion and spreading (see Mayr
& Reisigl, 1998).
Metaphors of racial, national, and ethnic
bodies and materiality are often also derived from
naturalizations; that is, from meteorologizations,
geologizations, and biologizations. To the latter
belong, in addition to the personifications already
mentioned (both nongendered and gendered), animalizations (of both sexes) and florizations.
Taking the example of German and Austrian discourses about migrants and racialized, national,
and ethnic minorities, the most frequent and stereotypical metaphors employed in the negative construction or identification of social actors and in the
negative predicational qualification of them, of their
migration, and of the alleged effects of immigration
are the following (see Bo ke, 1997, 2002; Gehard &
Link, 1991): natural disasters: immigration/migrants
as avalanches or flood disasters; dragging/hauling:
illegal immigration as dragging or hauling; water:
immigration/migrants as a water-course/current/
flood that has to be dammed; fire: alleged effects of
immigration/conflicts between racialized, nationalized, or ethnicized groups as a smoldering fire;
thermostatics: effects of immigration as pressure
within the pot and conflicts between racialized,
nationalized, or ethnicized groups as bubbling;
plants and fertile soil: migration and effects of migration as transplantation/repotting, uprooting, (alleged) causing of social conflicts as seeding; genetic
material: cultural and social traditions and heritage
as genetic material; growth/growing: increasing
immigration and increasing conflicts as growing; pollution and impurity: intergroup contacts, exchanges,
and relations as pollution and impurity; melting: intergroup contacts, relations, exchanges, and assimilation as melting; body: racialized, nationalized,
ethnicized groups are metaphorically ascribed collective (racial, national, ethnicity) bodies, outgroups
are metaphorized as foreign bodies or alien elements;
blood: immigration as bleeding white or bloodletting
of the imagined collective bodies, intergroup
relations as blood impurity; disease/infection:
immigration/migrants as an epidemic, intergroup
contacts and relations as an infection; animals/animal-owning: immigrants/minorities as parasites, as
attracted like the moths to a flame, as herded
together; war/fight/military: immigration as military
activity/invasion; goods/commodities and exchange
of goods: migration as import and export of workers/workforce, migrants as freight; food: good/
welcome immigrants/minorities versus bad/
unwelcome immigrants/minorities metaphorized as
the wheat that has to be separated from the chaff;
vehicle/boat/ship: effects of immigration as overcrowded boat; and house/building/door/gateway/

bolt: the ingroups (e.g., national) territory as
house or building, and stopping immigration as bolting the door. All of these metaphors are employed in
discourses about migrants and migration as implicit
argumentation schemes (i.e., as implicit and often
fallacious conclusion rules that serve to legitimise
discrimination; cf. Pielenz, 1993). The persuasive
function of these metaphors may be exemplified with
the metaphors of avalanches and flood. The two
metaphors suggest that If something (e.g., immigrants) is an avalanche or flood, it should be avoided
or prevented. The implicit argumentation function
of all the above-mentioned conceptual metaphors is
explicable by corresponding if-then schemes.
The suggestive force of metaphors can be so strong
as to motivate metaphorologists like Lakoff to contend that metaphors can kill (Lakoff, 1991). In view
of animalizing metaphors such as rat, parasite,
and vermin, referring to Jews during the National
Socialist dictatorship, and implying that if somebody
is a harmful vermin, rat, or parasite, he or she should
be exterminated, such a metonymical assertion does
not appear to be an exaggeration.
The tension between the preservation and transformation of political order, problem-solving, and
power relations becomes clearly manifest in the area
of metaphorical language. Just to offer a few hints:
static metaphors, metaphors of incorporation as well
as space and container metaphors seem to be favored
if questions of polity, policy, and politics are a matter
of consolidation and preservation, whereas destructive and (re)building metaphors, as well as dynamic
metaphors (e.g., metaphors of movement and
journey), indicate (intended) political transformation.
Metaphors with negative connotations implying
standstill are employed to evaluate preservation
negatively (e.g., fossilization of political structures).
Metaphors with negative connotations implying
political change are employed to evaluate transformation negatively (e.g., dismantling of political infrastructures).
Metonymies
Metonymies (from the Greek: renaming, name

change) are constituted by a shift involving two semantically (and materially, causally, or cognitively)
adjacent fields of reference: a name of a referent
stands for the name of another referent, which semantically (abstractly or concretely) adjoins the referent
of the name (cf. Morier, 1989: 749799; Lausberg,
1996: 292295; Groddeck, 1995: 233248; Plett,
2001: 98100).
Depending on the relationship between the two

neighboring conceptual fields, one can, among other
things, distinguish between the following metonymies: (1) the name of a product or effect stands for
the cause or author/creator (The racist book [standing for its author] provoked many enraged reactions.); (2) the name of a cause or author/creator
stands for the product or effect (The new Clinton
[standing for the new book authored by Clinton] is
a best-seller.); (3) the name of an object stands for the
user of the object (The cars [standing for their drivers] pollute and endanger the environment more and
more.); (4) the name of a container stands for the
containers content (Bush used to drink several
glasses [standing for the alcoholic liquid the glasses
were filled with] every day.); (5) the name of a place
(e.g., a state/country, town, city, district, or village)
stands for the person or persons living at the place
(Italy is not willing to accept more refugees.; Haider
says what Austria thinks.); (6) the name of a building
stands for the person or persons staying, working,
imprisoned, and so forth, there (The White House
decided to attack Iraq.; The concentration camp
Dachau was liberated in 1945 on May 2nd.); (7) the
name of a place stands for an action performed at the
place or an event located at the place (Auschwitz
must never ever happen again.); (8) the name of a
person stands for the country or state in which the
person is living (Cooperation is important because
we are too small to allow disharmony in vital areas of
our country., as the former Austrian foreign minister,
Wolfgang Schu ssel, said on 15 May 1995, in his
speech commemorating the fiftieth anniversary of the
signing of the international Austrian State Treaty.);
(9) the name of a person or of a group of persons who
have lived in the past stands for a person or a group of
persons who are living in the present (We already
fought against Prussians imperialism centuries ago.);
(10) the name of a time, time period, or epoch stands
for the persons living in the time, time period, or
epoch (The twentieth century has seen the most
extreme political developments in human history.);
(11) the name of an institution stands for (responsible) representatives of the institution (The government proposes a more restrictive asylum act.);
(12) the name of an institution stands for actions
performed within the institution or events that take
place in the institution (The Second Austrian Republic has almost always been described as a history of
successes.); and so on.
As one can see from examples 1, 3, 5, 6, 7,
10, 11, and 12, metonymies enable speakers and
writers within all of three political dimensions to
conjure away responsible, involved, or affected actors
(whether victims or perpetrators) or to keep them in
the semantic background. Example 9 illustrates how

a nationalist identification with the past manifests
itself by the usage of the historical we. Example
2 realizes the strategy of metonymic personalization,
which aims to stress the focus on a single politician.
Example 8 illustrates the strategy of metonymic concretizing and personifying (please note the doubled
tropic membership) that aims to promote political
identification and solidarity. Example 4 finally shows
a very conventional type of metonymy, which in the
specific context, could also be read as euphemistic
circumlocution of Bushs former alcoholism.
Metonymies of the type place name for persons
seem to be among the most frequent ones in political
discourses. They often realize the strategy of presupposing intragroup or intra-collective sameness or
similarity (see Wodak et al., 1999: 43). Such an assimilating or homogenizing function is, however, also
fulfilled by synecdoches and specific metaphors.
Synecdoches
Synecdoches (from the Greek: understanding one

thing with another) are turns of meaning within one
and the same semantic field: a term is represented by
another term, the extension of which is either semantically wider or semantically narrower (cf. Morier,
1989: 11591175; Lausberg, 1990b: 295298;
Zimmermann, 1989; Groddeck, 1995: 205220;
Plett, 2001: 9294). According to the direction
of representation, two types of synecdoches are
distinguished (cf. Plett, 2001: 9294).
The particularizing synecdoche is constituted by a
representative relation that consists of a semantically
broader concept standing for a semantically narrower
concept. The three main subcategories of this type of
synecdoches are pars pro toto (i.e., the part stands for
the whole: Hitler [representing Nazi Germany] was
finally defeated.), singularis pro plurale (i.e., the singular stands for the plural, forms a collective singular: The Swiss is industrious.), and species pro genus
(i.e., the species stands for the genus: The refugee
doesnt have a penny to her name [penny stands for
money].).
The generalizing synecdoche is established by a
semantically broader concept that represents a
semantically narrower one. The three principal subcategories of generalising synecdoches are totum pro
parte (i.e., the whole stands for the part: Denmark
is world champion. [this synecdoche is also a metonymy]), pluralis pro singulare (i.e., the plural stands for
the singular: We [representing a single person like
a queen or a states president; cf. pluralis maestatis

or pluralis modestiae] hereby enact a general amnesty.), and genus pro specie (i.e., the genus stands for
the species: Mankind has not learnt anything from
history.).
In political discourses, the particularizing synecdoche, just like the generalizing synecdoche, is a
means of referential annexation, assimilation, and inclusion. Particularizing synecdoches like the foreigner,
the Jew, and the American serve stereotypical generalization and essentialization, which refer in a leveling
manner to a whole group of persons. In languages like
German, Italian, and French, these are almost always
realized in their masculine grammatical form.
Lakoff & Johnson (1980: 38) have introduced as
a special conceptual synecdoche the synecdoche of
the controller for controlled type , in which leaders,
people in power, rulers, and so on stand for the
persons who are actually carrying out an action
(e.g. Bush bombarded Baghdad.; see also the first
synecdoche example). The use of this form of particularizing synecdoche is a central feature of all
monumentalist historiography (Nietzsche, 1985)
that hides social actors behind their leaders if speakers and writers judge them to be historically insignificant. Even if one does not assume, as do Lakoff and
Johnson (1980), that the synecdoche is a special form
of metonymy, it seems equally plausible to allocate
the type controller for controlled to the category of
metonymy.
The crucial role of synecdoches in polity, policy,
and politics has not yet been recognized adequately,
even though there are first attempts to grasp the
political relevance of this trope (see, for instance,
Palonen [1995], who analyses party names like The
Greens as synecdoches), and even though Kenneth
Burke stated more than 35 years ago that every act
of social representation and every theory of political
representation involves the trope of synecdoche, if
some part of the social body (either traditionally
established, or elected, or coming into authority by
revolution) is held to be representative of the society
as a whole (Burke, 1969: 508).
Burkes observation may serve as a starting point
for a general conceptualization of the relationship
between synecdoche and political representation.
Synecdochic representation is a basic political principle wherever more stands for less, or less stands for
more. Several factors determine whether or in what
way a political system of representation is organized
democratically: Who is entitled to be a political representative, and who decides on the question of who
will be a representative (active and passive right to

vote)? What does the procedure of political authorization to be a representative look like (direct or indirect democracy)? Under which conditions and when
can the relationship of representation be dissolved
or confirmed (representation for a set time)? To what
extent are representatives bound to those who are
represented (free or imperative mandate)?
There are different levels of political representation, which can be differentiated when looking at
political systems. On the highest level, there are to
be found, according to the political system, imperators, kings and queens (monarchy), states presidents
and chancellors (republic and democracy), or dictators (totalitarian dictatorship). Depending on the
level of representation, one and the same politician
is more or less representative. That is to say, at a
lower level of integration, a politician (e.g., a party
leader) may represent all members of a specific group
or political organization, whereas she or he just
represents a minority at a higher level of the system
(e.g., in the case that the party is a very small one).
These dynamics become very complex in modern
democracies, and all the above-mentioned determinants have to be taken into consideration by a tropological model of political representation (for more
details, see Reisigl, 2003: 252258).
Until now, an elaborated tropology of the political
does not yet exist, although a developed politolinguistic theory of tropes that scientifically and fruitfully
connects rhetoric, linguistic discourse analysis, and
political science is an urgent desideratum. Only such
a transdisciplinarian approach is capable of systematizing and explaining the many different political
and linguistic functions of tropic language in polity,
policy, and politics in a differentiated way that does
justice to the complex topic in question.
See also: Cognitive Science: Overview; Concepts; Creativ-
ity in Language; Figures of Speech; Irony; Metonymy;

Rhetoric, Classical; Thought and Language: Philosophical Aspects.
Bibliography
Bo ke K (1997). Die Invasion aus den Armenha usern
Europas. Metaphern im Einwanderungsdiskurs. In Jung
M, Wengeler M & Bo ke K (eds.) Die Sprache des Migrationsdiskurses. Das Reden uber Auslander in Medien,
Politik und Alltag. Opladen: Westdeutscher. 164193.
Bo ke K (2002). Wenn ein Strom zur Flut wird. Diskurslinguistische Metaphernanalyse am Beispiel der
Gastarbeiter- und Asyldiskussion in Deutschland und
sterreich. In Panagl O & Stu rmer H (eds.). 265286.
O
Bourdieu P (1991). Sozialer Raum und Klassen. Lecon

sur la lecon. Zwei Vorlesungen (2nd edn.). Frankfurt am
Main: Suhrkamp.
Burke K (1969). A rhetoric of motives. Berkeley: University
of California Press.
Chilton P (1996). Security metaphors: Cold War discourse from containment to common house. New York:
Lang.
Chilton P & Lakoff G (1995). Foreign policy by metaphor.
In Scha ffner C & Wenden A (eds.). 3759.
Chilton P & Ilyin M (1993). Metaphor in political discourse: the case of the common European house.
Discourse and Society 4, 731.
El Refaie E (2001). Metaphors we discriminate by:
naturalized themes in Austrian newspaper articles
about asylum seekers. Journal of Sociolinguistics 5/3,
352371.
Fairclough N (1995). Critical discourse analysis. The critical
study of language. London: Longman.
Fontanier P (1977). Les figures du discours. Paris:
Flammarion.
Gerhard U & Link J (1991). Zum Anteil der Kollektivsymbolik an den Nationalstereotypen. In Link J & Wu lfing W
(eds.) Nationale Mythen und Symbole in der zweiten
Halfte des 19. Jahrhunderts. Strukturen und Funktionen
von Konzepten nationaler Identitat. Stuttgart: Klett-Cotta.
1652.
Girnth H (1996). Texte im politischen Diskurs. Ein
Vorschlag zur diskursorientierten Beschreibung von
Textsorten. Muttersprache 1/1996, 6680.
Goodwin K (1990). Anthropomorphizing nationality.
Some Australian and African examples. In Zach W
(ed.) Literature(s) in English. New perspectives. Frankfurt am Main: Lang. 109128.
Goossens L (1995). Metaphtonymy: The interaction of
metaphor and metonymy in figurative expressions for
linguistic action. In Goossens L, Pauwels P, RudzkaOstyn S-V & Vanparys J (eds.) By word of mouth. Metaphor, metonymy and linguistic action in a cognitive
perspective. Amsterdam: Benjamins. 159174.
Groddeck W (1995). Rhetorik uber Rhetorik. Zu einer
Stilistik des Lesens. Basel: Stroemfeld.
Kienpointner M (1999). Metaphern in der politischen
Rhetorik. Der Deutschunterricht. Beitrage zu seiner Praxis und wissenschaftlichen Grudlegung 5/1999. 6678.
Klein J (2002). Weg und Bewegung. Metaphorische
Konzepte im politischen Sprachgebrauch und ein
frame-theoretischer
Repra sentationsvorschlag.
In
Panagl O & Stu rmer H (eds.) Politische Konzepte und
verbale Strategien Brisante Worter, Begriffsfelder,
Sprachbilder. Frankfurt am Main: Lang. 221235.
Lakoff G (1991). Metaphor in politics. An open letter
to the internet from George Lakoff. http://philosophy.
uoregon.edu/metaphor/lakoff-1.htm.
Lakoff G (1992). Metaphor and war: the metaphor system
used to justify war in the gulf. In Pu tz M (ed.) Thirty
years of linguistic evolution. Philadelphia: Benjamins.
463481.
Rhyme 605
Rhyme
K Hanson, University of California, Berkeley, CA, USA
Introduction
The Concise Oxford English Dictionary (COED)
defines rhyme as identity of sound between words
or verse-lines extending from end to last fully accented
vowel. The final words of the final couplets in (1) and
(2) from Shakespeares Sonnets are thus examples (1)
of a masculine rhyme because it involves just one
syllable, and (2) of a feminine rhyme because it
involves two (with three and even four syllable rhymes
also being possible, though unnamed by any gender):
(1) For thy sweet love remembred such
wealth brings,
That then I scorn to change my state with
kings. (29, 1314)
(2) But day doth daily draw my sorrows longer,
And night doth nightly make griefs length
seem stronger. (28, 1314)
The entry also lists the idiomatic phrase imperfect

rhyme, and illustrates it with the rhyme of love and
prove, also found as a final couplet in Shakespeares
Sonnets:
(3) But since he died and poets better prove,
Theirs for their style Ill read, his for his
love. (32, 1314)
Finally, the dictionary entry adds that in English prosody specifically, the identity cannot extend further
back from the end of the line than the last fully
accented vowel. However, that does occur in some
final couplets of Shakespeares Sonnets, as in (4):
(4) Then will I swear beauty herself is black,
And all they foul that thy complexion
lack. (132, 1314)
And rhymes that dont involve the last stressed vowel

also occur, as in (5), at least assuming the contemporary stress pattern on monument, a point to which we
return below:
(5) And thou in this shalt find thy monument,
When tyrants crests and tombs of brass
are spent. (107, 1314)
Even these few familiar examples show, then, that

while certainly conveying a common understanding
of rhyme, this definition excludes from its scope even
choices the most canonical English poet has made
to fulfill what is taken as a defining requirement of
the most canonical (and eponymous) form of English
poetry that a Shakespearean sonnet end with two

rhymed lines.
At the other end of the spectrum, the American
Heritage Dictionary, second college edition (AHD)
defines rhyme simply as correspondence of terminal
sounds of words or of lines of verse. This succeeds in
encompassing all the couplets of (1)(5) within its
scope, as well as more exotic examples we will see
below. But it does so partly by being vague: one might
well wonder, for example, what exactly is meant by
correspondence, or by terminal.
These two definitions and their respective hazards
represent the Scylla and Charybdis that any account
of rhyme must navigate exactly the same faced
by accounts of other linguistic forms such as relative
clauses or the intonational patterns of questions. The
description must be broad enough to encompass the
full range of structures having a plausible and significant claim to be instances of the form, and at the same
time it must be narrow enough to exclude imaginable
similar structures that in fact seem not to occur. In
addition to the usual methodological controversies
about how to achieve such a description, however,
an added difficulty for rhyme is that in one of its
central meanings at least that which takes rhyme
to be a characteristic of verse it is not simply a naturally occurring form but rather a product of highly
self-conscious, artful behavior designed to create
specific expressive effects.
Nonetheless, approaching the description of rhyme
as fundamentally a problem of linguistic form finds
substantial justification. To begin with, we can observe that both definitions agree that rhyme is a relationship between linguistic entities of some kind;
melodic figures passed from violins to viola to cello
in a string quartet may have identical or corresponding structures, but they do not rhyme, except perhaps
metaphorically. Any description of rhyme and its variant forms is thus inevitably at least partially a linguistic description. Still, the genuinely linguistic part
of the description might be limited; a description of
the crossword puzzle, for example, must necessarily
be partially linguistic, depending on reference to linguistic entities like the word, but it must also involve
other significant properties that have no role in language, such as whether a word is represented horizontally or vertically, or how many letters it is spelled
with. As it happens, however, what research there has
been so far into rhyme from an explicitly linguistic
point of view suggests that the relationship to language is quite a fundamental one; systematically,
growth in our understanding of language has yielded
corollary insights into rhyme.
606 Rhyme
In fact, even before getting into any details of such

descriptions, fundamental properties of language
allow rhyme to be delimited further in important
ways. First, the references to sound in both definitions
suggest that the relationship is a phonological one.
That distinguishes rhyme from at least one other
closely related verse form that has sometimes been
called rhyme, syntactic parallelism (Wimsatt, 1954).
At the same time, it also allows rhyme to be recognized in sign languages, which have phonology that is
not manifested in sound.
Furthermore, phonology itself is held to be organized
along two quite distinct dimensions rhythm, involving syllables, stress, phrasing, and so forth, and melody,
the qualitative characteristics that define individual
segments, such as the features that make an [s] different
from a [z] or a [t] or an [n], or an [i] different from an [e]
(Liberman and Prince, 1977; Goldsmith, 1979).
Given this distinction, it seems clear that rhyme necessarily involves a relationship based on linguistic melody: simple and border, for example, have similar
phonological structures in virtue of their rhythmic
shapes, yet they do not rhyme in English; we will see
below that this is not merely a prescription, even an
intuitively acceptable one, but rather a description
based on an empirical claim that poets do not seem
to use such pairings to fulfill their rhyme schemes.
While the family of verse forms based on rhythmic
repetition has long been given the common name
of poetic meter, rhyme clearly joins such forms as
alliteration, assonance, and consonance in forming a
family crucially based on melodic relationships. These
are sometimes collectively called sound-patterning,
but because it is not clear in what sense rhythm is
not sound, except in the sense referred to earlier in
which phonology in general is not sound, these forms
might usefully be given an alternative common name
such as poetic harmony as proposed by Gostusky
(1969), because they involve structured relationships
among the component parts of linguistic melodies.
Both definitions also take rhyme to be a relationship
between ends. It is this that distinguishes them from at
least one of these other poetic harmonies, alliteration,
which by definition relates melodies at beginnings,
such as the [d]s at the beginnings of the words day,
daily, draw (and perhaps also doth) in (2) above.
Finally, we can assert without any argument beyond common experience, because it will be developed much more fully below, that the nature of the
relationship, somewhere between the COEDs identity and the AHDs correspondence, involves some
kind of likeness like all verse forms, rhyme is a form
of repetition.
These considerations give, then, as a starting point, a
somewhat revised version of the capacious definition
given by the AHD of rhyme as minimally a relationship of likeness between final linguistic melodies. So
far, however, this revision does little to address the
vagueness of the original. In fact, we will see that the
best way to do that seems to be not to define rhyme
more narrowly, but rather to articulate the specific
sources of the variation found in rhymes across languages and across individual poets practices. In a
useful article modestly purporting merely to bring
some order to the chaos of overlapping terminology
used for rhyme and other harmonic forms, Rickert
(1978) identifies three distinct sources of variation:
where the language involved in the relationship is
located in a text, exactly which segments in those
locations are constrained by the form, and exactly
what degree of similarity those constraints require.
The first two specifications, we will see, can jointly
answer the question of what exactly is meant by
final in our definition, and the third will answer the
question of what exactly constitutes a likeness, our
compromise between identity and correspondence.
Thus, the range of possible answers to these three
questions can generate a typology of rhyme, and it
is around them that the following discussion is
organized.
What makes these questions of particular interest
to linguists is the theoretical question to which we
referred briefly earlier, the question of the extent to
which the range of possible answers might itself be
constrained by linguistics. For example, building on
the perspective of Jakobson (1960), Kiparsky (1973,
1987) articulated a substantive hypothesis, already
latent in much inquiry into rhyme, that the structures
in terms of which poetic forms are defined are precisely those that also figure in grammar. In particular,
he suggested that rhyme bears significant resemblances to reduplication. If this is correct, we would
expect answers to our questions to make reference
to linguistic entities such as word boundaries, syllable nuclei or back vowels, but not, for example, to
important words, or even, as is often assumed to be
the case, to spelling. Of course, the issue is complicated by the fact that the hypothesis defines other
poetic forms too, such as metrical feet and lines,
which are likewise linguistic in the sense of being
defined in reference to linguistic entities but are not
part of ordinary language and to which descriptions
of rhyme must plainly make reference. Perhaps
more cautiously, then, we can say, especially in light
of the artistry with which rhyme is created as noted
above, that answers to our questions that refer to linguistic entities define rhymes that are natural, either
directly, referring to structures of ordinary language,
or indirectly, referring to specifically poetic forms
that themselves are natural in this way; poets may
Rhyme 607
or may not also create rhymes that are defined in

other ways and that are thus definably artificial. In
any case, this interesting theoretical question is far
from settled, nor is there even that much in the way
of rigorous description of rhyme available against
which to test it. On the contrary, even for such a
well-documented poetic tradition as that of English,
in many ways the study of rhyme is in its infancy, rich
with open research questions.
Finally, a word is in order about the methodology
that has shaped linguistic inquiry into rhyme. First,
the question of what does or does not rhyme is
essentially a question about an individual poets competence in exactly the sense that term has within linguistics: knowledge that governs intuitions about
well-formedness, whether conscious or unconscious,
codified or not. That competence may also be complex in the same ways as other linguistic knowledge
can be, encompassing differences according to
genre and purpose or period of composition, just as
speakers may control different registers, and their
languages may include various dialects. Generalizations about rhyme must therefore always be understood as essentially claims about commonalities
across individual competences.
Second, the best and usually only evidence for a
poets intuitions about rhyme is his or her actual
practice. As Kiparsky (1973) noted, answering the
question of what counts as a rhyme for a poet is really
only possible to the extent that it is possible to identify a rhyme scheme a pattern in which rhymes recur
in some arrangement that is sufficiently predictable
that it is possible to assert that two items that seem
rather different nonetheless must rhyme in the poets
estimation because the scheme he or she has created
calls for them to, and because there is no reason to
think the scheme has simply been abandoned. For
example, it is the claim that the Shakespearean sonnets discussed earlier share a rhyme scheme of
ababcdcdefefgg that motivates the insistence that
rhyme must be defined for Shakespeare in such a
way as to account for the fact that for him, all of the
pairings in (1)(5) are evidently rhymes. Kiparskys
term for this is that the rhyme is obligatory as opposed to optional, the same distinction drawn between the more traditional terms structural and
ornamental. It is always worth bearing in mind,
however, that obligatory is also relative to different
languages, periods, and genres; as Fabb (2002) noted,
the interpretation of a scheme as obligatory is essentially a pragmatic inference that a poet has in fact
committed him or herself to producing a rhyme at a
certain point, and as such it is both contextual and
fallible in all kinds of ways. In fact, the questions of
what rhyme schemes occur in poetry and how exactly
we come to perceive them are thus fascinating and

open in themselves, but also in many ways beyond the
bounds of linguistics proper, even as assumptions
about them underpin linguistic investigation.
In what follows, then, we return our attention
to the narrower question of the form of rhyme and
some recent attempts to offer precise descriptions
of particular rhyme practices, clarifying their relationship to aspects of phonological structure and exploring parallel phenomena in grammar itself. The
illustrations will be drawn mostly from English because the data from English can be presented with the
most confidence, but it should be emphasized that
key concepts are drawn from work on not only
English but also Finnish, Irish, Turkish, Hebrew,
French, German, Dutch, Romanian, and many other
languages, as indicated in the bibliography.
The Location of Rhymes

We have said that rhyme is by definition final, a
melodic likeness between ends; but our first question
is, ends of what? As we will see, the question does not
necessarily have a simple answer: the hierarchical
organization of both linguistic rhythm and verse
structure makes available a rich variety of constituents whose ends are in principle available for defining
rhymes, and are in fact combined in various and
sometimes complex ways by different poets.
In the canonical case of perfect rhyme illustrated in
(1), the rhymes occur at the ends of metrical lines. By
definition, a metrical line has a rhythmic structure
just as all language does as part of its phonological
structure that consists of its parse into such entities as
syllables (s), feet (f), words (o), phrases (F), etc., in
what is inspiringly but somewhat confusingly called
its metrical phonology (see Metrical Phonology).
But a line of metrical verse also has a (poetic) metrical
structure, a parse that shows its phonological rhythmic structure to satisfy whatever additional constraints define the meter. The metrical parse is thus
closely related to the phonological parse, but also in
principle is distinct from it. One way of modeling
meter that makes the distinction particularly vivid is
through a metrical template that defines some abstract rhythmic structure together with a set of correspondence rules that define mappings of the rhythmic
structure given by the phonology into that template
(see Meter). For iambic pentameter, for example, the
meter of the example in (1), the template has been
argued to be that in (7), consisting of five rising feet,
each composed of a weak position (W) followed by a
strong one (S), with these feet grouped in turn ambiguously into two rising cola, which are grouped finally
into a line:
608 Rhyme
syllable with an unstressed one in what is sometimes

called a tangential rhyme can be seen when scanned
as in (9) to follow from exactly the same characterization as rhymes like those in (8) as located in the
final syllables, which are those in the final strong
positions:
(9) And thou in this shalt find thy monument,
w
s w s w s
w
s w s
When tyrants crests and tombs of brass are spent.
w s w
s w s
w s w
s
The iambic pentameter of Shakespeares Sonnets

includes the constraint that there is one syllable in
each position, and (8) shows the language of the lines
in (1) mapped into the template in accordance with
that constraint and others (such as a constraint that
stressed syllables of polysyllabic words like remembered be mapped into strong metrical positions). The
template is shown fully for this first example, but
abbreviated thereafter; and the rhythmic structure of
the language is likewise partially represented just this
once to give the basic idea:
The interest of this for rhyme is that when language

is mapped into the template in this way, and so defined as metrical, it is in a sense also defined as having
two ends, each in its turn definable at a variety of
levels: there is the end of the template the final
strong position, foot, colon, etc. as well as the end
of the language mapped into it its final syllable (s),
foot (f), word (o), phonological phrase (F) etc.
In lines like those in (8), the end of the language
coincides with the end of the template in what is
called a masculine ending, whence the term masculine rhyme for rhymes between lines of this type. It is
therefore possible to locate the rhymes in either the
final syllables of the lines, or the final strong positions
of the lines, without particularly needing to attend to
any potential distinction between the two. But this is
by no means the case for all rhymes. Even for these
simple masculine rhymes, note that once the role of
meter in locating rhymes is properly articulated, it is
no longer necessary or even correct for accounts of
rhyme in verse to locate rhymes in terms of a last
accented vowel, at least not for Shakespeares Sonnets. Rhymes like that in (5) that rhyme a stressed
In fact, it seems quite possibly generally true that

stress does not matter to well-formedness in rhyme
directly, but only indirectly through its contribution
to meter whence the confinement of rhyme to
melodic likeness suggested earlier. Note that this predicts that vowel length, too, might well be generally
irrelevant to well-formedness in rhyme. But at the
same time, further research might well turn up a
practice that exploits this distinction between metrical and rhythmic structure in genuinely locating
rhymes in syllables containing last accented vowels,
rhyming spent in (9) for example with entity, but not
with monument.
Meters also allow non-coincidence of the end of
the language of a line and the end of a template. In a
metrical version of phonological extrametricality, for
example, the iambic pentameter of Shakespeares
Sonnets allows an optional additional unstressed syllable (<>) line-finally, as in the scansion of (2) in
(10), traditionally called a feminine ending, and
whence the term feminine rhyme for a rhyme between two lines of this type:
(10) But day doth daily draw my sorrows longer,
w s w s w s w s w s <>
And night doth nightly make griefs length seem
w s
w s w s
w
s
w
stronger.
s <>
In the common case illustrated here, the location

of the likeness that defines the rhyme begins in the
syllable in the final strong position, as in the rhymes
in (1), but continues through the end of the language
of the line.
In English, it is most common for masculine lines to
rhyme with masculine lines and feminine with feminine, but these are not the only possibilities. There are
also what are called androgynous rhymes in which
the two line types are mixed. Shakespeares Sonnets
include at least one possible rhyme of this type, and
his plays include other clear examples. In the bs of
the abab quatrain in (11a), scanned in (11b), the
syllable in the strong position of the feminine line
does not figure in the rhyme: as it does not in the
couplet in (12a), scanned in (12b).
Rhyme 609
(11a) For when these quicker elements are gone
In tender embassy of love to thee,
My life being made of four, with two alone
Sinks down to death, oppressed with
melancholy; (45: 6,8)
(11b) In tender embassy of love to thee,
w s w s w s w s w s
Sinks down to death, oppressed with melancholy;
w
s w s
w s
w s w s <>
(12a) Quite overcanopied with luscious woodbine,
With sweet musk-roses and with eglantine
(A Midsummer Nights Dream [italicized]
2.2.251-2)
(12b) Quite overcanopied with luscious woodbine,
w s wsw s
w s w
s <>
With sweet much-roses and with eglantine
w
s
w
s w s
w s w s
Instead, the syllable that the final one of the masculine line is like is the extrametrical syllable in the
feminine line. The meter thus defines what counts as
the line, but the end of it that is relevant to rhyme is
the end of the language mapped into the template and
not the end of the template itself.
Conversely, in another type of androgynous rhyme
sometimes called apocopated (but perhaps overdue
to be revised to something more like drag queen
rhyme so that the nameless adventurous structures
in (12), those feminine lines behaving as if they were
masculine, can be recognized as drag king rhymes),
it is the extrametrical syllable that does not figure in
the rhyme. The likeness holds instead between the
syllables in the final positions, regardless of what
follows them. An example (for which I am indebted
to Michael Ferguson) is the rhymes of the mixed
iambic pentameter and tetrameter couplets of James
Merrills The Octopus as in (13), scanned in (14) (on
the assumption that the iambic pentameter of this
poem differs from Shakespeares Sonnets in allowing
two unstressed syllables in a single weak position,
(see Meter):
(13) There are many monsters that a glassen surface
Restrains. And none more sinister
(14) There
are many monsters that a glassen surface
|{z}
w
s w s w s w s w s <>
Restrains. And none more sinister
w s
w s
w sws
Finally, these possibilities suggest that even

when both lines are feminine, some poets might
rhyme only the syllables in the final metrical positions, or only those that are final in the language, or
mix the two.
Nor is extrametricality the only source of complexity in locating rhymes. The foregoing are examples of
what is called end-rhyme, which in spite of rhyme
being always an end phenomenon, is not redundant,

because it refers specifically to the rhymes being located at ends of lines, when they might be at ends of
other constituents. Internal rhyme, another special
case noted by the COED, refers generally to rhymes
in which at least one partner is not final in the line.
Internal rhyme can be as systematic as end rhyme,
and also as complex. One relatively straightforward
example might be the rhyme between sed and med in
the third line of the stanza of the Middle English
Cuckoo song in (15), which can be characterized as
a rhyme of the final syllable (or position) in the line
with the final syllable (or position) in the first colon of
the same line:
(15) Sumer is i-cumen in
Lhude sing, cuccu!
Groweth sed and bloweth med
And springth the wude nu.
Sing, cuccu!
Medieval Irish included a variety of stanzaic forms

defined by both end rhyme schemes and internal
rhyme schemes, including systematic rhymes between
the ends of certain lines and the ends of the first cola
or even in some cases the first words of the succeeding
lines, as in the rhymes of amra and fagla, or elta and
Cerca in (16), examples from Murphy (1961) for
which I am indebted to Eve Sweetser (and in which
the nature of the likeness defining the rhymes will be
discussed more fully below):
(16) Anmchaid Osraigi amra
ca ini fagla flathrgi
drecon bruthmar bruithe elta
mac Con Cerca cathmli.
Some of these internal rhymes, it should be noted,

deepen the question of exactly what is a line. Why, for
example, should a line like the third one in (15) be
counted as one long line with internal rhyme rather
than as two short ones with end rhyme? Should
rhyme perhaps be seen as capable of creating lines,
rather than simply marking ends of ones defined by
meter? After all, meter is not the only way of defining
a linguistic entity that has an end. Is rhyme ever based
on the ends simply of phrases, or of sentences? In
modern poems that define lines only visually through
page-layout, do rhymes occur at ends whose definitions bypass considerations of language and meter
entirely? The scope for complex relationships arising
through rhyme from diverse ways of defining ends,
while not without limits, is far greater than standard
accounts of rhyme have tended to acknowledge. Poets
have found these complexities, and linguists and
critics are just beginning to catch up: Shoptaw
(2004), for example, has recently argued that some
610 Rhyme
of Dickinsons poems involve counter-lineation, in

which visual and metrical lines pointedly and significantly diverge, and interlocking rhyme patterns are
defined over both.
And of course, internal rhymes are not necessarily
systematic at all; in a quatrain like that in (17) from
another of Shakespeares Sonnets, the most that can
be said of its abundant internal rhymes is that they are
final in the words involved:
(17) Will will fulfill the treasure of thy love,
Aye, fill it full with wills, and my will one.
With things of great receipt with ease we prove,
Among a number one is reckoned none:
(135, 58)
In sum, then, we can say that rhyme is indeed

a relationship of likeness between ends, and that it
is only possible to speak of a rhyme scheme if it is
possible to specify which ends in some systematic way.
At the same time, the rich range of constituents made
available by language and meter let alone by nonlinguistic structures such as page layout provide a
surprisingly complex and varied range of possible
structures whose articulation and delimitation raise
many interesting open questions for a theory of
rhyme.
The Domains of Different Rhyme Types

The question of what is final involves not just where
what is final actually ends, but also where what is final
starts or put another way, over exactly what domain,
measured how, likeness must hold in order to define
a rhyme. Here again, there is a wide range of imaginable possible answers, and the specifically linguistic
character of rhyme emerges in the extent to which the
rhymes poets actually use are best characterized
through reference either to linguistic constituents or
to metrical constituents closely modeled on them.
In the example in (1) of a perfect masculine
rhyme, the language that is alike matches neatly the
portion of the syllable actually called the rhyme on
analogy to the verse form. One fairly standard description of the internal structure of the syllable is
that represented in (18), with the first segment in the
nucleus being the most sonorous one, hence in most
English syllables the first vowel. On this analysis,
brings and kings would have the syllable structures
in (18a,b), and hence identical syllable rhymes:
However, because in English at least, poets mix

such rhymes quite freely with feminine ones, and in
the latter the entire syllable succeeding the one in the
strong position normally matches as well, as in (2, 10),
this resemblance turns out to be rather less useful to a
description of rhyme than might be expected. Rather,
what the internal structure of the syllable seems to
delimit is the beginning of the domain over which
likeness holds (Holtman, 1996). The form of rhyme
in which what is matched runs from the nucleus of
some specified final syllable through its coda and
on to some specified end, then, is the form known as
full rhyme. In many ways, taking both these monosyllabic masculine and polysyllabic feminine rhymes
into account, what the domains resemble is the portion of a foot from its head on, with the metrical
strong position defining that head; but given the
complexity we have just seen regarding potential distinctions between metrical and rhythmic structure,
the question of exactly what the domain of likeness
is in polysyllabic rhymes clearly requires more research and careful formulation, especially involving
traditions in which polysyllabic rhymes are the norm,
such as Irish, as seen in (16) above.
Another significant form of rhyme involves less
than the syllable rhyme, and that is slant rhyme. In
slant rhyme, syllable nuclei are irrelevant to wellformedness; only final consonants need bear any likeness to one another, as in the a rhymes of spoke and
Lake in the first aabb stanza of Yeatss Under Ben
Bulben in (19):
(19) Swear by what the sages spoke
Round the Mareotic Lake
That the Witch of Atlas knew
Spoke and set the cocks a-crow.
Here, too, there are interesting questions about

how exactly to describe the form linguistically. First,
it should be noted that it is only final consonants and
not entire syllable codas that are alike in slant rhyme,
at least in English; the form provides no evidence for
the coda as a constituent. For example, in the stanza
from Yeatss Sailing to Byzantium in (20), the abababcc rhyme scheme is satisfied by the rhyme of
unless and dress with magnificence:
(20) An aged man is but a paltry thing,
A tattered coat upon a stick, unless
Soul clap its hands and sing, and
louder sing
For every tatter in its mortal dress,
Nor is there singing school but studying
Monuments of its own magnificence;
And therefore I have sailed the seas
and come
To the holy city of Byzantium.
Rhyme 611
A second interesting question involves rhymes such

as that of the second couplet in (19), in which no final
consonants are present. In English, there seem to be
two distinct traditions. One is represented by Yeatss
practice, in which the vowels must be either both
back, as in the rhyme of knew and crow in (19), or
else both front as in, for example, the rhyme of I and
utterly in Easter 1916 (1315). Since non-low tense
vowels in English always end their articulation with
glides sharing their features [ back], Yeatss rhyme
seems to recognize the presence of final glides in
apparently open syllables and to require that they be
alike. In the other tradition, however, vowel-final
rhymes observe no such constraint, as in, for example, Plaths rhyme of go and lie in the abcacb scheme
of All the dead dears:
(21) And to sanctuary: usurping the armchair
Between tick
And tack of the clock, until we go,
Each skulled-and-crossboned Gulliver
Riddled with ghosts, to lie
Deadlocked with them, taking root as cradles rock.
The latter case is, as observed by Kiparsky (1973),

exactly parallel to the well-known pattern in Germanic alliteration, whereby initial consonants must
match each other to satisfy the form, but initial
vowels can match any other vowel; it is also in harmony with claims that no final glides are present in
underlying representations of simple tense vowels
(Chomksy and Halle, 1968). The simultaneous existence of the tradition represented by Yeats, however, suggests that there are open questions about
the best way to model both practices and to represent the structures of the relevant syllables in
English.
A final form of melodic repetition that is arguably a
form of rhyme is assonance. Assonance is the reverse
of slant rhyme: in assonance, nuclei must bear a likeness to one another, but the consonants following
them need not. One argument for treating assonance
as a form of rhyme is that, although not as common
as full rhyme or slant rhyme, it is used to instantiate
schemes defined over line ends, as in some medieval
French romances or in Dylan Thomass Fern Hill,
where assonances like green and leaves or climb and
eyes and light instantiate an abcddabcd scheme
(Zwicky, 1976):
(22) Now as I was young and easy under the
apple boughs
About the lilting house and happy as the
grass was green,
The night above the dingle starry,
Time let me hail and climb,
Golden in the heyday of his eyes,
And honored among wagons I was prince

of the apple towns
And once below a time I lordly had the
trees and leaves
Trail with daisies and barley
Down the rivers of the windfall light.
(19)
Here, the existence of phonological tiers that

represent only vowels may be key to understanding
assonance as a form of rhyme that can genuinely be
said to be based on language that is final in some
sense.
In sum, the domains over which the likenesses that
define rhymes hold are not necessarily constituents.
Conversely, certain linguistic constituents whose existence has been posited do not necessarily define
rhyme types. At the same time, however, it seems
that the demarcation of the domains of likeness is a
matter of linguistic structure, either directly, or indirectly through metrical structure, insofar as their
beginnings are the beginnings of well-defined constituents, rather like the domains posited by prosodic
morphology (McCarthy and Prince, 1996).
It is also worth noting that it is sometimes suggested that rhymes are defined not just by required
likeness, but also sometimes by required difference;
the COED, for example, as mentioned earlier explicitly asserts that in English the likeness must not
continue (moving back in the line) beyond the nucleus
of the first syllable to include its onset, and some
accounts of rhyme by linguists have attempted to
model this constraint (Holtman, 1996; Yip, 1999).
However, for English at least, this seems better understood as an evaluative prescription (possibly reflecting rejections of French rhyme in which such
likenesses are actually highly valued) than as a systematic practice that a description of rhyme must
account for; in (4) earlier, for example, we have already seen a counterexample, and Oras (1955) has
argued that rhymes of just this type have been used
in English throughout the tradition.
Partial Likenesses in Rhymes

Perhaps the richest question of all that rhyme poses
for linguistics is that of what, exactly, counts as a
sufficient likeness for a rhyme. It has long been recognized that in any of the earlier forms, sometimes the
likeness that defines a rhyme involves mere similarity
rather than complete identity. In one of the most
influential articulations of the relevance of linguistics
to the study of rhyme, Zwicky (1976) pioneered the
explicit description of differences allowed in rhymes
through reference to distinctive features, which was
pursued further by Golston (1997) and others.
Focusing on the rhymes of English rock lyrics,
612 Rhyme
Zwicky describes a form he calls rock rhyme, which

is a full rhyme in the sense given earlier of requiring
likeness from the nucleus in the final strong position,
but which alongside other special structures discussed further below allows matched segments to
differ by a single distinctive feature in any position.
Thus in the abcbdefe scheme of Bob Dylans Masters
of war in (23); in the first stanza, desks rhymes with
masks, whose vowels [E] and [] differ in the latter
being [low], while in the ababab scheme of Ballad
of Hollis Brown, in the third stanza in (24), sleeve
rhymes with breathe, whose [v] and [] differ in the
latters not being [labial]. And both the nuclei and
the coda consonants differ (by [ back] and [ labial],
respectively) in the rhyme of guns/bombs in (23):
(23) Come you masters of war
You that build all the guns
You that build the death planes
You that build the big bombs
You that hide behind desks
I just want you to know
I can see through your masks
(24) Your babys eyes look crazy
Theyre a-tuggin at your sleeve
Your babys eyes look crazy
Theyre a-tuggin at your sleeve
You walk the floor and wonder why
With every breath you breathe
Crucially, Zwicky argues that the rhyme of sleeve

and breathe in (24) is not assonance, because across
the whole song, the final consonants are not free as
they are in genuine cases of that form, as in (22), but
rather subject to the feature-based constraint on similarity he identifies. By the same reasoning, the rhyme
of desks and masks in (23) is not slant rhyme, because
across the song, the nuclei do not vary without limit
they vary only within the scope of the posited constraint. Put another way, full rhyme, slant rhyme, and
assonance are defined just as discussed above by the
different domains they impose constraints on final
nuclei and what follows, final consonants alone, or
final nuclei alone, respectively the question of
whether those constraints require complete identity
is in principle a separable one. If the constraints do
not require complete identity, but do require some
similarity that is statable in a systematic way, that
becomes an argument for the inclusion of even differing segments within the domain of required likeness,
and the rhyme may be termed partial.
As Malone (1993, 1987) has pointed out, the general phenomenon Zwicky noted of the likenesses that
define rhymes being satisfied by partial rather than
complete identity is in fact not at all confined to rock
lyrics but appears throughout the English tradition as
well as in the rhyme traditions of other languages.

When the practices of individual poets and traditions
are considered separately, however, it is not as clear
that Zwickys definition of similarity as a singlefeature difference is ever the best way to characterize
the operative constraints. Rather, these seem to take a
wide (though not limitless) variety of possible forms
that submit to a wide (though not limitless) variety of
explanations involving linguistic structure in a somewhat different way.
Perhaps the most familiar source of rhymes involving similarity rather than strict identity with a linguistic explanation is differences in pronunciation. An
audience in one dialect may perceive only similarity
where a poet composed a strict identity in another.
For example, the assonance of Fern Hill in (22) is
realized at one point by the pairing barns/calves,
in which the vowels are different in many American
dialects but not in Thomass Welsh dialect. A particularly complex version of this question arises when
describing the poets dialect is itself a problem of
historical reconstruction. For example, the question
of whether Shakespeares vowels are really different
in quality in his rhymes of prove/love in (3) or even in
stress in monument/spent in (5) is simultaneously a
question about his rhyme practice and about his dialect, in that any reconstruction of his pronunciation
must take into account the possibility that partial
rather than complete identity may have sufficed for
a rhyme, just as much as any account of his rhyme
practice must take into account the possibility that
these rhymes were strictly identical in his dialect
(Ko keritz, 1953; Hanson, 2002). For it is clear that
even partial rhymes that have their origins within
the tradition in changed pronunciations cannot necessarily be claimed to have been complete at their
moment of composition in the dialect of every poet
who has used them; whatever their pronunciation in
Shakespeares dialect may have been, rhymes like
love/prove persisted in being composed in English
long after their vowels qualities had undeniably diverged. It has sometimes been argued that such partial
rhymes have continued to be used productively simply because the original usage has established them as
conventional pairings, an approach commonly taken
to both the many slight differences occurring, for
example, in the vowels of English full rhymes and
the differences in the consonants in Zwickys rock
rhymes, modeled as they often are on blues sung by
musicians in whose dialects certain final consonants
were deleted (Kiparsky, personal communication).
But such accounts of historical origin do not preclude
the possibility that the form in which they persist is
not arbitrary pairings but grammatical generalizations (Hanson, 2002). In general, then, the influence
Rhyme 613
of dialect variation over both time and place remains

an important open question in the study of virtually
every poets rhymes.
Another explanation for partiality in rhymes of
linguistic interest involves the question of at what
level of representation the relevant correspondence
holds. An influential hypothesis put forth by Kiparsky
(1968, 1972), that identities that appear only partial
on the surface actually are complete in underlying
representations, has been shown to explain rhyme
patterns in a variety of languages (see, for example,
Malone (1982, 1988) on Turkish or Malone (1983)
on Hebrew). This hypothesis might even be illustrated for English by a simple case such as the
rhyme of monument/spent in (5). Even if, as suggested
earlier, the difference in the stress of their final syllables is irrelevant to the well-formedness of the rhyme,
there remains the issue that, at least in contemporary
pronunciation of these words, there is also a difference in the quality of their vowels, [e] and [E]. But
when the final syllable of monument is stressed as in
mo`nume ntal, the vowel that surfaces as e in monument surfaces as E instead; hence we may hypothesize
that that vowel is E underlyingly, and that that might
therefore be the vowel figuring in the rhyme.
There is no question, however, that all partial likenesses in rhyme can be reanalyzed as complete in
either of these ways. On the contrary, in some traditions, the partial likenesses that are allowed in rhyme
are even codified. In Irish, for example, consonants
are grouped into classes, any member of which can
count as the same as the others for purposes of rhyme.
One class, for example, is the k class of the voiceless
stops [p], [t], and [k] (spelled c); and another is the
l class of voiced fricatives together with what are
called weak sonorants, hence [v] (sometimes spelled
b or bh), [] (sometimes spelled d or dh, [g]
(sometimes spelled g or gh), [b] (sometimes spelled
m or mh) and [m], [n], [l] and [r]. In the internal
rhymes mentioned in (16) earlier, then, which are
polysyllabic rhymes, amra (a[br]a) and fagla
(fa[gl]a) rhyme because their medial consonants belong to the same succession of classes, as do those of
elta (e[lt]a) and Cerca (ce[rk]a). A question of interest
for the relationship of rhyme to linguistics is that of in
what sense classes of this type are natural classes from
a phonological point of view and what role they have
in the languages ordinary grammar. In Irish, these
classes used in rhyme do figure in the grammar, for
example, as either targets or results of rules of lenition
(Eve Sweetser, personal communication; Malone,
1987; Murphy, 1961.
And even without being codified in such a way,
phonologically systematic partial rhymes emerge in
other traditions. In English, for example, surprisingly
many poets systematically allow consonants that differ

in voice to nonetheless count as rhyming (Frawley,
1984). It is sporadic and historically debatable in
some early texts, as in the particularly haunting
rhyme of glass/was from Shakespeares Sonnets:
(25) Then were not summers distillation left
A liquid prisoner pent in walls of glass,
Beautys effect with beauty were bereft,
Nor it nor no remembrance what it was. (5, 12)
But it is pervasive and undeniable in many other

poets practice, such as Pinskys slant rhyme translation of Dantes Inferno, which regularly matches not
only [s] with [z] but also [t] with [d], [f] with [v] and
[S] with [Z], as in his stanzas of aba bcb cdc .. terza
rima in (26) (Hanson, 2003):
(26) a. A different path from this one would be best
For you to find your way from this feral place,
He answered, seeing how I wept. This beast,
The cause of your complaint, lets no one pass
Her way but harries all to death. Her nature
Is so malign and vicious she cannot
appease (I 7175)
b. It seemed it might never stop; behind it a train
Of souls so long that I would not have thought
Death had undone so many. When more than one
I recognized had passed, I beheld the shade
Of him who made the Great Refusal, impelled
By cowardice: so at once I understood (III 4752)
c. Who set it there? He said: It should be clear:
Over these fetid waves, you can perceive
What is expected if this atmosphere
Of marsh fumes doesnt hide it. Bow never drove
Arrow through air so quickly as then came
Skimming across the water in a little
skiff (VIII 1116)
d. You will be brought to shore by another way.
My master spoke then, Charon, do not rage:
Thus is it willed where everything may be
Simply if it is willed. Therefore, oblige,
And ask no more. That silenced the grizzled jaw
Of the gray ferryman of the livid marsh, (III 7681)
Other common partial matches in English involve

nasals that differ in place, as in Dylans rhyme of guns
and bombs in (23), or another of Pinskys in (27):
(27) If it should please you, stop a moment here.
Your way of speaking shows that you
were born
In the same noble fatherland: there where
I possibly have wrought excessive harm.
This sound erupted from a coffer of stone
I drew back toward my guide in my alarm. (X 2227)
And as in Irish, these equivalences in rhyme invite

phonological inquiry into why these particular partialities should be used. The discounting of voice in
rhyme in English, for example, can be related to the
neutralization of voice distinctions in codas in English
that leads to alternations like that between [s] and [z]
614 Rhyme
in the plural (as in cat[s] but dog[z]) or [t] and [d] in

the past (as in bark[t] but meow[d]) (Hanson, 2003).
This, in turn, can be related to a more general diminution of the perceptibility of the voicing contrast in
final position, as suggested by Steriade (forthcoming).
Even in English, however, not all poets allow the same
partially similar segments to count as alike in rhyme;
Dickinson, for example, allows rhymes of obstruents
that differ in place, a pairing not used by Pinsky that
suggests that different poets partition the phonological space in different ways.
In sum, some partiality in what counts as a likeness
is actually the rule rather than the norm in at least
some poetic traditions, and, it should be emphasized,
is available in all rhyme forms, for example, in full
rhyme, just as much as in slant rhyme or assonance.
Such partiality may have a variety of sources, but in
all cases, the question of how the specific partialities
poets use are related to linguistic structure seems a
fruitful and worthy avenue of study on a poet-by-poet
basis in a wide range of languages.
always final, and also always coronals. This seems

to hold true for many other English poets as well;
allowing a final extra [s] or [z] is widespread, with
some poets also allowing an extra [t] or [d], and in
some complex cases involving feminine endings, possibly even an extra [n]. Why should such structures
count as rhymes? Golston related the phenomenon to
observations by Borowsky (1986) and others that
English syllable rhymes normally permit only three
segmental positions (with long vowels counting as
two), thus allowing for example sneak ([Ink]) or
ink [iNk] but not *sn[iNk]. Final coronals, however,
are allowed in excess of this constraint, as in fiend
([ind]) or fierce ([irs]), suggesting that the final coronals may be external to the syllable rhyme, perhaps
incorporated instead as an appendix as in (30):
Subsequence
A final complexity in rhyme also noted by Zwicky for
English rock rhymes, but seeming to have far wider
applicability as well as significant linguistic interest,
is subsequence. On Zwickys description, subsequence is the presence in one rhyme partner of a
segment not matched in the other; an example again
from Masters of War is the final [z] in eyes not present
in its rhyme partner fly in the abcbdefe scheme in
(28), while the extra [t] in fleetst compared to its
partner sweets in (29) shows a similar structure in
an abab quatrain from Shakespeares Sonnets:
(28) You that never done nothin
But build to destroy
You play with my world
Like its your little toy
You put a gun in my hand
And you hide from my eyes
And you turn and run farther
When the fast bullets fly
(29) Make glad and sorry seasons as thou fleetst,
And do what eer thou wilt, swift-footed Time,
To the wide world and all her fading sweets:
But I forbid thee one most heinous
crime: (19, 58)
Subsequence seems, however, to be somewhat

more constrained than Zwickys description allows,
and in ways that point to interesting possible explanations for it in relation to syllable structure
(Golston, 1997). Golston observed that in Dylans
rhymes, subsequent segments can be argued to be
Furthermore, among all the coronals, [s/z] has been

argued to be peripheral to syllable structure on different grounds, including its being excepted from the
universal constraints that the sonority of segments
should decrease or at least remain level across a
coda as in the [rs] of fierce, rather than increasing as
it does in, for example, the [ks] of sphinx (Harris,
1994). The proper analysis of these linguistic phenomena is debated; but in any case it is clear that
subsequence in rhyme provides an at least partially
parallel phenomenon whose description and explanation awaits research.
Resemblances to Reduplication
We have up to this point considered rhyme as a verse
form that both as a whole and in individual poets
practices seems to bear systematic relationships to the
phonology of ordinary language. Nothing in the description of the phenomenon we have developed so far,
however, confines rhyme to verse; in fact, neither does
either dictionary definition with which we began, because both allowed for rhyme simply at ends of
words. It is, in fact, perfectly reasonable to speak
even more generally of words whose syllables rhyme,
whether in the manner of full rhyme as in picnic or
claptrap, or slant rhyme in riffraff, or even of apocopated rhyme in, say, a phrase like a very married pair.
The relationship of rhyme to ordinary phonology
that has probably been of the most interest to linguists, however, is not sporadic rhymes, whether
Rhyme 615
intentional or not, or even the influence of rhyme

on vocabulary development; but rather the fully systematic grammatical process of reduplication.
Kiparskys suggestion, with which we began, that
the linguistic structures that figure in verse are fundamentally the same ones that figure in grammars, offered parallels with reduplication as its evidence
(specifically, parallels between alliteration in Germanic verse and Gothic past tense formation). A few
linguists have developed the parallel in fruitful and
extensive ways, including Holtman (1996), Yip
(1999), and, with reference specifically to language
acquisition, Inkelas (2003). It seems quite likely that
understanding such issues as the polysyllabic rhythmic structures onto which melodies are copied in
reduplication, or melodic simplifications implemented in copying, will illuminate answers to the basic
question of why poets count what they do as rhymes.
Conclusion
The interest of rhyme for linguistics may be a matter
of its form alone, but the consequences of description and explanation of its forms reach beyond linguistics into literary criticism and theory. In the
interpretation of poetry, one of the most vexing questions is that of the nature of the contribution forms
make to meaning, or more generally to aesthetic effect not just the legitimacy of arguments from one to
the other, but even the legitimacy of positing the
existence of such forms to start with, especially in
the face of the evident variation they exhibit. While
linguistic investigation of rhyme by no means settles
these philosophical questions, attention to it expands
enormously our awareness of the range of phonology
that seems available to give rise to aesthetic effects,
both in modulation across obligatory rhyme schemes
(Lanz, 1931) and in relationships forged by optimal
ones (Wimsatt, 1954). The project should thus be of
interest to literary critics and linguists alike.
See also: Meter; Metrical Phonology; Poetry: Stylistic Aspects; Semantics of Prosody.
Bibliography
Berube M S (ed.) (1982). The American heritage dictionary.
(2nd college edn.). Boston: Houghton-Mifflin Co.
Borowsky T J (1986). Topics in the lexical phonology of
English. Ph.D. diss., University of Massachusetts.
Brogan T V F (1981). English versification 15701980: a
reference guide with a global appendix. Baltimore: Johns
Hopkins University Press.
Chomsky N & Halle M (1968). The sound pattern of
English. Cambridge, MA/New York: Harper and Row.
Fabb N (1997). Linguistics and literature. London:

Blackwell.
Fabb N (2002). Language and literary structure. Oxford:
Blackwell.
Frawley W (1984). A note on the phonology of slant
rhyme. Language and Style 17(1), 4447.
Goldsmith J (1979). Augosegmental phonology. New York:
Garland.
Golston C (1997). The geometry of rhyme. Ms., CSU
Fresno.
Gostusky D (1969). The third dimension of poetic expression or language and harmony. Musical Quarterly 55,
372383.
Hanson K (2002). Vowel variation in English rhyme. In
Minkova D & Stockwell R (eds.) Studies in the history of
English: a millenial perspective. The Hague: Mouton.
207229.
Hanson K (2003). Formal variation in the rhymes of
Robert Pinskys The Inferno of Dante. Language and
Literature 12(4), 309337.
Harris J (1994). English sound structure. Oxford: Blackwell.
Holtman A (1996). A generative theory of rhyme: an optimality approach. Ph.D. diss., Universiteit Utrecht.
Utrecht: LEd (OTS dissertation series).
Inkelas S (2003). Js rhymes: a longitudinal case study of
language play. Journal of Child Language 30, 557581.
Jakobson R (1960). Closing statement: linguistics and poetics. In Sebeok T (ed.) Style in language. Cambridge,
MA: MIT Press and J. Wiley & Sons. 350377.
Kiparsky P (1968). Metrics and morphophonemics in the
Kalevala. In Gribble C (ed.) Studies presented to
Professor Roman Jakobson by his students. Cambridge,
MA: Slavica. Reprinted in Freeman D (ed.) (1968). Linguistics and literary style. New York: Holt, Rinehart and
Winston. 137148.
Kiparsky P (1972). Metrics and morphophonemics in
the Rigveda. In Brame M (ed.) Contributions to generative phonology. Austin: University of Texas Press.
171200.
Kiparsky P (1973). The role of linguistics in a theory of
poetry. Daedalus 102(3), 231244.
Kiparsky P (1987). On theory and interpretation. In Fabb
N et al. (eds.) The Linguistics of Writing. New York:
Methuen. 185198.
Ko keritz H (1953). Shakespeares pronunciation. New
Haven: Yale University Press.
Lanz H (1931). The physical basis of rime: an essay
in the aesthetics of sound. Stanford: Stanford University
Press.
Liberman M & Prince A (1977). On stress and linguistic
rhythm. Linguistic Inquiry 8(2), 249336.
Malone J (1982). Generative phonology and Turkish
rhyme. Linguistic Inquiry 13, 550553.
Malone J (1983). Generative phonology and the metrical
behavior of u- and in the Hebrew poetry of medieval
Spain. Journal of the American Oriental Society 103(2),
369381.
Malone J (1987). Muted euphony and consonant matching
in Irish verse. General Linguistics 27(3), 133144.
616 Rhyme
Sykes J B (ed.) (1976). The concise Oxford dictionary of
current English (6th edn.) (1st edn., 1911). Oxford: The
Clarendon Press.
Stallworthy J (1983). Versification. In Allison A W et al.
(eds.) The Norton Anthology of poetry, 3rd edn. New
York: W. W. Norton. 14031423.
Steriade D (2003). Knowledge of similarity and narrow
lexical override. BLS 29.
Steriade D (Forthcoming). The phonology of perceptibility
effects: the P-map and its consequences for constraint
organization. In Hanson K & Inkelas S (eds.) The nature
of the word: essays in honor of Paul Kiparsky,
Cambridge: MIT Press.
Wimsatt W K (1954). One relation of rhyme to reason. In
Wimsatt W K (ed.) The verbal icon: studies in the meaning of poetry. Lexington, KY: Kentucky University Press.
153168.
Yip M (1999). Reduplication as alliteration and rhyme.
GLOT International 4, 17.
Zwicky A (1976). Well, this rock and roll has got to stop.
Juniors head is hard as a rock . Cls 12, 676697.
Malone J (1988). Underspecification theory and Turkish

rhyme. Phonology 5, 293297.
Malone J (1993). Partial rhyme. Hellas 1, 107122.
McCarthy J & Prince A (1996). Prosodic morphology
1986. RuCCS-TR32. Rutgers Center for Cognitive
Science.
Murphy G (1961). Early Irish metrics. Dublin: Royal Irish
Academy.
Oras A (1955). Intensified rhyme links in The Faerie
Queene: an aspect of Elizabethan rhymecraft. The Journal of English and Germanic Philology 55, 3960.
Perloff M (1970). Rhyme and meaning in the poetry of
Yeats. The Hague: Mouton.
Pinsky R (1998). The sounds of poetry: a brief guide. New
York: Farrar, Straus and Giroux.
Preminger A (ed.) (1974). The Princeton encyclopedia of
poetry and poetics. enlarged edition. Princeton: Princeton
University Press.
Rickert W E (1978). Rhyme terms. Style 12, 3546.
Shoptaw J (2004). Listening to Dickinson. Representations 86, 2052.
Rhythm
rhythmic structure. In spoken language, rhythm
also comes up in conjunction with rhythmic adjustments of different kinds (destressing, stress shift,
addition of stress). The pattern of regular alternation
is thought of as prototypically rhythmic, but the notion of rhythm is more inclusive than that in linguistics, especially when we extend the focus to verse
metrics. In meter, less symmetric patterns, such as
those found in the verse of Sappho, would also qualify as rhythmic, provided they occur with some
regularity. Thus, rhythm could be defined as an
arrangement of prominent and nonprominent prosodic elements that can be perceived as a regular
pattern.
Rhythmic structure is hierarchical and it has been
represented both as tree structure and grid structure
in the phonological literature. In the following grid
structure, hierarchy is coded as different heights of
the columns, reflecting level of prominence. This
would be a typical rhythm of a song or nursery
rhyme (example from Hayes (1995: 26)).
T Riad, Stockholm University, Stockholm, Sweden

Definition
Rhythm is the perception of a pattern of alternation
between prominent and nonprominent prosodic elements. In a linguistic context, rhythm is mostly discussed in the domains of stress patterns and verse
metrics. Rhythm is deeply involved in the shape of
stress systems, stress being probably the only phonological category that has a rhythmic distribution
(Hayes, 1995: 25). The rhythmic character of stress
distribution is evident, for example, in six-syllable
words, in which stress will tend to occur on every
other syllable (sBsBssBss or sBssBssBs). No language
will assign stress to three adjacent syllables in such
a word (*sssBsBsBs or ssBsBsBss). Furthermore, it
is natural to view the categories of linguistic stress
(the binary prosodic feet) and their distribution
in words as part of the linguistic manifestation of
(1) x
x
x
x
This
he
knickthis
x
x
old
played
knack
old
x
x
x
man
knickpadman
dy
x
x
knack
wack,
came
x
x
x
x
he
on
give
rol-
your
x
x
played
my
dog
ling
x
x
x
one,
thumb,
bone,
home.
x
x
with
616 Rhyme
Sykes J B (ed.) (1976). The concise Oxford dictionary of
current English (6th edn.) (1st edn., 1911). Oxford: The
Clarendon Press.
Stallworthy J (1983). Versification. In Allison A W et al.
(eds.) The Norton Anthology of poetry, 3rd edn. New
York: W. W. Norton. 14031423.
Steriade D (2003). Knowledge of similarity and narrow
lexical override. BLS 29.
Steriade D (Forthcoming). The phonology of perceptibility
effects: the P-map and its consequences for constraint
organization. In Hanson K & Inkelas S (eds.) The nature
of the word: essays in honor of Paul Kiparsky,
Cambridge: MIT Press.
Wimsatt W K (1954). One relation of rhyme to reason. In
Wimsatt W K (ed.) The verbal icon: studies in the meaning of poetry. Lexington, KY: Kentucky University Press.
153168.
Yip M (1999). Reduplication as alliteration and rhyme.
GLOT International 4, 17.
Zwicky A (1976). Well, this rock and roll has got to stop.
Juniors head is hard as a rock . Cls 12, 676697.
Malone J (1988). Underspecification theory and Turkish

rhyme. Phonology 5, 293297.
Malone J (1993). Partial rhyme. Hellas 1, 107122.
McCarthy J & Prince A (1996). Prosodic morphology
1986. RuCCS-TR32. Rutgers Center for Cognitive
Science.
Murphy G (1961). Early Irish metrics. Dublin: Royal Irish
Academy.
Oras A (1955). Intensified rhyme links in The Faerie
Queene: an aspect of Elizabethan rhymecraft. The Journal of English and Germanic Philology 55, 3960.
Perloff M (1970). Rhyme and meaning in the poetry of
Yeats. The Hague: Mouton.
Pinsky R (1998). The sounds of poetry: a brief guide. New
York: Farrar, Straus and Giroux.
Preminger A (ed.) (1974). The Princeton encyclopedia of
poetry and poetics. enlarged edition. Princeton: Princeton
University Press.
Rickert W E (1978). Rhyme terms. Style 12, 3546.
Shoptaw J (2004). Listening to Dickinson. Representations 86, 2052.
Rhythm
rhythmic structure. In spoken language, rhythm
also comes up in conjunction with rhythmic adjustments of different kinds (destressing, stress shift,
addition of stress). The pattern of regular alternation
is thought of as prototypically rhythmic, but the notion of rhythm is more inclusive than that in linguistics, especially when we extend the focus to verse
metrics. In meter, less symmetric patterns, such as
those found in the verse of Sappho, would also qualify as rhythmic, provided they occur with some
regularity. Thus, rhythm could be defined as an
arrangement of prominent and nonprominent prosodic elements that can be perceived as a regular
pattern.
Rhythmic structure is hierarchical and it has been
represented both as tree structure and grid structure
in the phonological literature. In the following grid
structure, hierarchy is coded as different heights of
the columns, reflecting level of prominence. This
would be a typical rhythm of a song or nursery
rhyme (example from Hayes (1995: 26)).
T Riad, Stockholm University, Stockholm, Sweden

Definition
Rhythm is the perception of a pattern of alternation
between prominent and nonprominent prosodic elements. In a linguistic context, rhythm is mostly discussed in the domains of stress patterns and verse
metrics. Rhythm is deeply involved in the shape of
stress systems, stress being probably the only phonological category that has a rhythmic distribution
(Hayes, 1995: 25). The rhythmic character of stress
distribution is evident, for example, in six-syllable
words, in which stress will tend to occur on every
other syllable (sBsBssBss or sBssBssBs). No language
will assign stress to three adjacent syllables in such
a word (*sssBsBsBs or ssBsBsBss). Furthermore, it
is natural to view the categories of linguistic stress
(the binary prosodic feet) and their distribution
in words as part of the linguistic manifestation of
(1) x
x
x
x
This
he
knickthis
x
x
old
played
knack
old
x
x
x
man
knickpadman
dy
x
x
knack
wack,
came
x
x
x
x
he
on
give
rol-
your
x
x
played
my
dog
ling
x
x
x
one,
thumb,
bone,
home.
x
x
with
Rhythm 617
Rhythmic structure strives toward even spacing at all

intervals of repetition. Thus, feet tend to yield alternating rhythm, as in Example (2). Phrasal stresses,
too, tend to be evenly spaced, even though the number of unstressed syllables vary (see later, Example
(9)). If the distance becomes too long, new, rhythmically motivated stresses will occur. Rhythmic structure also exhibits synchronization of beats at all
levels, as in Example (1) (downward implication).
Thus, a syllable receiving phrasal stress will also
carry a word stress and will be the head of a foot.
Representationally, this means that high grid marks
are always on the top of a column. These properties
are all testable through ordinary experience, i.e.,
tapping in response to rhythm and sensing what is
natural and what is not.
Resolution of clash and lapse involves the deletion

and addition of stress beats. In principle, the combination of those two processes will also describe
the process of stress shift. So-called iambic reversal
(the rhythm rule, stress retraction) of English is one
example of this:
Eurhythm
As seen, the stress shifts to the next prominence on

its own level, rather than just to the next syllable,
which would create an ill-formed grid structure. Example (6) also illustrates the fact that rhythmic adjustment need not be triggered by a clash in the
narrowest sense, since in this example, an unstressed
syllable comes between the clashing ones.
The picture that emerges is one in which the linguistic feet encode the basic rhythmic property of
alternation, and linguistic rules adjust the distribution
of stress in connected speech. The upshot is an
approximation of equally spaced prominences. However, the feet and the adjustment rules are not enough
to achieve even spacing.
Alternating rhythm is unmarked, and unmarked

rhythm is optimal in language. Optimal rhythm, or
eurhythm, is to some extent attained by virtue of the
canonical shapes of linguistic feet as iambic (. x) or
trochaic (x .). Sequences of iambs or trochees will
inevitably instantiate alternating rhythm:
(2) Iambic:
Trochaic:
(. x)
ss
(x .)
ss
(. x)
ss
(x .)
ss
(. x)
ss
(x .)
ss
(. x)
ss
(x .)
ss
Eurhythm is attained also by adjustment processes of

various kinds. There are two prototypical nonrhythmic contexts (so-called arrhythmies) to which languages typically show a reaction: clash and lapse.
Clash is the situation whereby two prominences are
adjacent at some level. The specification of what
counts as a clash and a lapse may vary according to
language (Nespor and Vogel, 1989). The typical reaction to clash is the reduction of the lesser stress, as in
the following cases from Swedish:
(3) "kvar-%skatt
!
"kvar-skatt
residual tax
tull-%pack-%hus-%karl-s-%lag!
"tull-pack-%hus-karl-s-%lag
customs packing house mens team
In the second example, the resulting rhythm ("tullpack-%hus-karl-s-%lag) is regularly alternating, by virtue of every other stress being retained. Lapse the
situation when the distance between stresses is too
great is resolved by the insertion of a rhythmic
prominence somewhere between the stresses, as in
the following example from Italian (Nespor and
Vogel, 1989: 86):
(4) Gia nni non ce lo da
! Gia nni non ce` lo da
Gianni does not give it to us.
(5) th`rteen men ! th`rteen men
Clash resolution by stress shift respects the hierarchical level at which stress clash occurs. This can
be brought out in a grid representation (Prince,
1983):
Rhythm in the Speaker and Listener

Rhythm does not arise simply as a function of the
sheer distribution of phonological prominences in a
linguistic structure. It also arises by virtue of the
execution of the prosody of a linguistic structure,
wherein the distribution of stresses can be made
more even by the regulation of speech tempo. Furthermore, the interpretation of auditory impressions goes
some way toward creating rhythm. The tendency
to adjust the distance between stresses is typical of
languages such as Swedish or English, and it is
supremely clear in verse metrics:
(7) x
. x
. x . . x
Humpty Dumpty sat on a wall
x
. x
. x
. . x
Humpty Dumpty had a great fall
(8) x
. . x . .
x . .
x
. .x . . x .
Sjung, o guidinna om vreden som brann hos peliden Akilles
Sing, O goddess, the anger of Achilles son of Peleus
x . x . . x . . x . . x. . x .
Olycksdiger till tusende kval for akajernas soner
that brought countless ills upon the Achaeans.
618 Rhythm
In each of these verse samples, a nursery rhyme and

the first two lines of the Iliad in a Swedish translation,
the stresses are equally spaced, despite the fact that the
number of unstressed syllables between them varies.
We find the same phenomenon in spoken language,
but here the isochrony is more approximate. In the
following example, taken from Engstrand (2004:
205) there are four stress groups, containing radically
different numbers of syllables:
(9)
.
x . . .
.
x .
.
x
.
x
s | sss s
s |
s s s | s
s | s
Och | tjejerna kom och | skrek att det | brann som | fan
And the-girls came and shouted that it burnt like hell.
In the actual pronunciation of this utterance, the

groups are of roughly equal length, which means
that the syllables of the first group are pronounced
much faster than are those of the third and fourth.
Regulations in speech tempo also affect the very application of rhythmic adjustment rules. Indeed, in
slow and careful speech, adjacent stresses might not
induce the sense of clash, and in fast speech, some
syllables may be suppressed altogether.
The degree of isochrony varies depending on structural properties of the language, such as phonotactics,
quantity system, and gradation (Strangert, 1985). Indeed, there is a traditional notion that languages
could be categorized as syllable timed or stress
timed (e.g., Abercrombie, 1967), the idea being that
some languages would strive toward equal distance
between syllable peaks (Italian, Spanish), whereas
other languages would strive toward equal distance
between stresses (German, English). In an influential
paper, Dauer (1983) showed that rhythm in the languages in her sample were all stress based, acoustically, and that they come differently close to isochrony.
In fact, the evidence for acoustically even spacing is
quite weak. Perceptually, however, things are different. Perception may give a rather strong sense of
isochrony even when the physical correlate is quite
uneven, and it might also allow for the impression of
syllable timing. The findings of Dauer show that it is
possible to retain stress as the single phonological unit
in language that has a rhythmic distribution. Syllables
do not form the base for rhythm, and neither do other
linguistic categories. Also, it is clear that phonological
rules pertaining to stress apply so as to create even,
rather than uneven, rhythm, as we have seen.
Grouping: Iambic/Trochaic Law

Beside its role in the rhythm of spoken language,
perception is also crucially implicated in the derivation of linguistic feet that is, grouping, from basic
input to rhythm. The empirical finding among stress
systems in languages is that iambic systems tend to
have long syllables in the prominent position of the

foot, whereas trochaic systems tend to have short
syllables in the same position. Stated differently,
there seems to be a bias to have uneven quantity in
the two positions of an iamb, but even quantity in the
positions of a trochee. There is a perceptual base for
this, known in phonology as the iambic/trochaic law,
which builds on the insights of psychological experiments on rhythmic grouping (see Hayes, 1995: 79,
and references therein):
(10)
!
!
( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
If the input varies in intensity, people will tend to

group the inputs in units with initial prominence,
viz., trochees. If the input varies in quantity, the
inputs will be grouped as iambs, with final prominence. Beyond foot structure, there are phonological
rules of lengthening and shortening in iambic and
trochaic systems, respectively, that work in the direction of optimizing the pattern in accordance with
Example (10) (see Mester, 1994).
Meter
In spoken language, rhythmic properties are essentially automatic. In poetic meter, prosodic properties of
the language become codified. Unmarked properties
may be enhanced, as in the isochrony of the nursery
rhymes shown in Examples (1) and (7). However, in
many styles of verse, it is considered crude to overstate the rhythmic pattern in the manner done in
nursery rhymes, and rhythm is instead signaled
more subtly. One exponent for this in the English
tradition is blank verse, in which none of the metrical
positions actually requires of the syllable filling it that
it be either stressed or unstressed. Similarly, meter can
be consciously made to deviate from the unmarked.
Indeed, any verse tradition containing more than one
meter must have less marked meters in it. Classical
Greek meter is traditionally considered a quantitative
system, based on long and short syllables, but it
explores arrhythmic properties such as clash and
lapse in the construction of meter. Thus, for dactylic
meter, the essential rhythmic property is one clash per
metron, and for iambic meter it is one lapse per
metron. Interestingly, clash and lapse are not very
strong constraints in the language, and it would
seem that this is what makes it possible to explore
these arrhythmies in the verse tradition (Golston and
Riad, 2000). For languages without stress, e.g.,
Japanese, we expect that they should have very little
in the way of rhythmic phenomena in both language
and verse. The verse types used in Japanese (haiku,
tanka), which count moras but exhibit no sensitivity
Rhythmic Alternations 619
to prominence, certainly seem to support this conclusion (Kozasa, 2000).

See also: Meter; Metrical Phonology; Poetry: Stylistic
Aspects; Rhyme; Rhythmic Alternations; Semantics of
Prosody.
Bibliography
Abercrombie D (1967). Elements of general phonetics.
Edinburgh: Edinburgh University Press.
Dauer R M (1983). Stress-timing and syllable-timing
reanalyzed. Journal of Phonetics 11, 5162.
Engstrand O (2004). Fonetikens grunder. Lund: Studentlitteratur.
Golston C & Riad T (2000). The phonology of Classical

Greek meter. Linguistics 38(1), 99167.
Hayes B (1995). Metrical stress theory. Principles and case
studies. Chicago: University of Chicago Press.
Kozasa T (2000). Moraic tetrameter in Japanese poetry.
Manuscript, University of Hawaii at Manoa.
Mester R A (1994). The quantitative trochee in Latin.
Natural Language & Linguistic Theory 12, 161.
Nespor M & Vogel I (1989). On clashes and lapses.
Phonology 6, 69116.
Prince A (1983). Relating to the grid. Linguistic Inquiry 14,
19100.
Strangert E (1985). Swedish speech rhythm in a crosslanguage perspective. (Acta Universitatis Umensis 69).
Stockholm: Almqvist & Wiksell.
Rhythmic Alternations
D Zec, Cornell University, Ithaca, NY, USA
Introduction
Rhythmic alternation is a principle of linguistic organization that operates at the level of sound structure
and is rooted in the cognitive organization of auditory
events. As an organizing cognitive principle, rhythmic
alternation grants cohesion to temporal spans, by
virtue of propelling the regularity of event sequencing
from past to future. This general cohesive role also
characterizes rhythmic alternation in linguistic constituencies. Rhythm has a specific organization in
terms of beat structure, detected in language as well
as in other cognitive domains that are crucially dependent on the time dimension. According to studies
of the rhythmic structures in music (Cooper and
Meyer, 1960; Lerdahl and Jackendoff, 1983), as
well as investigations of linguistic rhythm (Liberman,
1975; Liberman and Prince, 1977; Hayes, 1995),
auditory stimuli are construed as beats, and a sequence of beats is further construed as a sequence of
groups of beats with recurring group internal structure. In its turn, group internal structure is manifested
as greater prominence conferred on one of the group
members, typically the first or last, thus yielding an
alternating pattern.
Studies of the cognitive properties of rhythm have
reported on the general conditions for the emergence
of rhythmic alternation (Fraisse, 1978; Woodrow,
1951). According to Woodrow (1951: 1232), [b]y
rhythm, in the psychological sense, is meant the perception of a series of stimuli as a series of groups of
stimuli. The successive groups are ordinarily of similar
pattern and experienced as repetitive. Trochaic

grouping tends to be perceived when the sequence
consists of equidistant auditory events alternating in
loudness, whereas iambic grouping emerges when
alternating sounds are longer in duration, or are followed by a longer pause. The optimal temporal interval for rhythmic grouping is relatively small, to
facilitate the perception of a group as a whole, that
is, as a single perceptual event. These findings about
the general cognitive properties of rhythm elucidate
the nature of rhythmic patterning in language.
The Scope of Rhythmic Alternation

in Language
Rhythmic alternation in linguistic constituents is
most directly evidenced in alternating stress patterns.
Phonological theories have captured the basic aspects
of rhythm by representing both grouping and prominence (Halle and Vergnaud, 1987; Halle and Idsardi,
1995; Hayes, 1995; Prince, 1990). Thus, in Example
(1), each syllable of M`ssissppi, which has an alternating stress pattern, corresponds to a beat (line (a)),
with the first and third beats more prominent than the
second and fourth (line (b)), and the third beat more
prominent than the first (line (c)). Beats are represented by the symbol x, groups are identified by parentheses, and degrees of prominence are captured by
the height of the beat column.
(1) (c)
(b)
(a)
(
(x
(x
x)
M ` s s i s s
x
x
(x
p p
)
)
x)
i
However, rhythmic alternation in language is constrained in several ways. Whereas domains such as
to prominence, certainly seem to support this conclusion (Kozasa, 2000).

See also: Meter; Metrical Phonology; Poetry: Stylistic
Aspects; Rhyme; Rhythmic Alternations; Semantics of
Prosody.
Bibliography
Edinburgh: Edinburgh University Press.
Dauer R M (1983). Stress-timing and syllable-timing
reanalyzed. Journal of Phonetics 11, 5162.
Engstrand O (2004). Fonetikens grunder. Lund: Studentlitteratur.
Golston C & Riad T (2000). The phonology of Classical

Greek meter. Linguistics 38(1), 99167.
studies. Chicago: University of Chicago Press.
Kozasa T (2000). Moraic tetrameter in Japanese poetry.
Manuscript, University of Hawaii at Manoa.
Mester R A (1994). The quantitative trochee in Latin.
Natural Language & Linguistic Theory 12, 161.
Phonology 6, 69116.
Prince A (1983). Relating to the grid. Linguistic Inquiry 14,
19100.
Strangert E (1985). Swedish speech rhythm in a crosslanguage perspective. (Acta Universitatis Umensis 69).
Stockholm: Almqvist & Wiksell.
Rhythmic Alternations
D Zec, Cornell University, Ithaca, NY, USA
Introduction
Rhythmic alternation is a principle of linguistic organization that operates at the level of sound structure
and is rooted in the cognitive organization of auditory
events. As an organizing cognitive principle, rhythmic
alternation grants cohesion to temporal spans, by
virtue of propelling the regularity of event sequencing
from past to future. This general cohesive role also
characterizes rhythmic alternation in linguistic constituencies. Rhythm has a specific organization in
terms of beat structure, detected in language as well
as in other cognitive domains that are crucially dependent on the time dimension. According to studies
of the rhythmic structures in music (Cooper and
Meyer, 1960; Lerdahl and Jackendoff, 1983), as
well as investigations of linguistic rhythm (Liberman,
1975; Liberman and Prince, 1977; Hayes, 1995),
auditory stimuli are construed as beats, and a sequence of beats is further construed as a sequence of
groups of beats with recurring group internal structure. In its turn, group internal structure is manifested
as greater prominence conferred on one of the group
members, typically the first or last, thus yielding an
alternating pattern.
Studies of the cognitive properties of rhythm have
reported on the general conditions for the emergence
of rhythmic alternation (Fraisse, 1978; Woodrow,
1951). According to Woodrow (1951: 1232), [b]y
rhythm, in the psychological sense, is meant the perception of a series of stimuli as a series of groups of
stimuli. The successive groups are ordinarily of similar
pattern and experienced as repetitive. Trochaic

grouping tends to be perceived when the sequence
consists of equidistant auditory events alternating in
loudness, whereas iambic grouping emerges when
alternating sounds are longer in duration, or are followed by a longer pause. The optimal temporal interval for rhythmic grouping is relatively small, to
facilitate the perception of a group as a whole, that
is, as a single perceptual event. These findings about
the general cognitive properties of rhythm elucidate
the nature of rhythmic patterning in language.
The Scope of Rhythmic Alternation

in Language
Rhythmic alternation in linguistic constituents is
most directly evidenced in alternating stress patterns.
Phonological theories have captured the basic aspects
of rhythm by representing both grouping and prominence (Halle and Vergnaud, 1987; Halle and Idsardi,
1995; Hayes, 1995; Prince, 1990). Thus, in Example
(1), each syllable of M`ssissppi, which has an alternating stress pattern, corresponds to a beat (line (a)),
with the first and third beats more prominent than the
second and fourth (line (b)), and the third beat more
prominent than the first (line (c)). Beats are represented by the symbol x, groups are identified by parentheses, and degrees of prominence are captured by
the height of the beat column.
(1) (c)
(b)
(a)
(
(x
(x
x)
M ` s s i s s
x
x
(x
p p
)
)
x)
i
However, rhythmic alternation in language is constrained in several ways. Whereas domains such as
620 Rhythmic Alternations
music demonstrate a cognitive capacity for groupings

at multiple hierarchical levels of rhythmic structure
(Lerdahl and Jackendoff, 1983), rhythmic constituency in language is fairly shallow. Hierarchical rhythmic
organization in language is highly restricted and
mainly follows from the hierarchical organization of
the linguistic constituents that serve as potential
domains for rhythmic patterning. Likewise, whereas
the cognitive capacity for manipulating groups with
more than two members, and with more elaborate
internal organization, is evidenced in domains other
than language (Fraisse, 1978), rhythmic patterns in
language tend toward simple, binary groupings, as in
Example (1). In fact, precisely because such alternating patterns are relatively simple, they can equally
well be represented by the so-called grid structure,
which lacks explicit grouping (Prince, 1983; Hayes,
1984, 1995; Selkirk, 1984; Gordon, 2002), as illustrated in Example (2). Representations in Examples
(1) and (2) have identical prominence structures yet
differ in how the regular alternation of prominent
beats is formally executed. In Example (2), an alternating pattern is secured by invoking principles such
as eurhythmy, a requirement for an even distribution
of stresses, which should be neither too close to each
other nor too far apart. That is, neither adjacent
stressed syllables, or clashes, nor adjacent unstressed
syllables, or lapses, are permitted:
(2) (c)
(b)
(a)
x
x
x
x
x
x
x
M ` s s i s s p p i
Ternary groupings, while rare, have also been evidenced

among alternating stress patterns. In these systems,
stressed syllables alternate with intervals of two
unstressed syllables. Phonological analyses either admit
ternary groupings into the set of possible group types
(Halle and Vergnaud, 1987; Levin, 1988), or posit
binary groups separated by an additional unstressed
syllable (Hayes, 1995), or posit grid structures with a
special license for lapses (that is, adjacent unstressed
syllables) (Elenbaas and Kager, 1999; Gordon, 2002).
Patterns of Rhythmic Alternation in

Word Phonology
Rhythmic alternation belongs to the range of possible
phonological patterns for expressing prominence
relations within words. Rhythmic alternation is
most directly manifested in stress systems, which provide an abundance of relevant cases and set the scope
within which this organizational principle is manifested in human language. However, rather than
being a general property of word stress, rhythmic
alternation characterizes only those stress systems

that permit more than one stress within the word
domain. Moreover, alternating patterns are not restricted to stress, and are encountered in other word
prominence systems as well. Groupings in either
stress or non-stress alternating patterns are subject
to the same principles, rooted in the cognitive aspects
of rhythmic alternation.
Alternating Patterns of Stress
Crucial in the classification of stress systems is the

prosodic unit that serves as the basic rhythmic beat.
Units that naturally emerge in this role are the mora
and the syllable, both members of the prosodic hierarchy (Hayes, 1981, 1995; Halle and Vergnaud,
1987; Kager, 1995; and the references cited in these
works). The syllable is the optimal candidate, due to
its stable internal constituency and the high sonority
of its nucleus. There are, however, two independent
issues relevant for beat selection: which unit serves as
a phonetically realized beat, and which serves as a
beat for the purposes of computing an alternating
pattern. Empirical evidence points at the syllable as
the phonetically realized beat, but stress in general,
and its alternating patterns in particular, can be computed with either the mora or the syllable taken as the
basic beat. The two classes of stress systems, syllabic
and moraic, are further subclassified into those with
trochaic (left-dominant) or iambic (right-dominant)
grouping. This yields four types of stress systems.
Three of those, syllabic trochees, moraic trochees,
and moraic iambs, are amply exemplified in Hayes
(1995) and Kager (1995). Syllabic iambs present a
special case and are addressed here.
Regular rhythmic alternation is most readily exhibited in systems of syllabic trochees. An alternating
pattern can be computed either from left to right, as
in Example (3), or from right to left, as in Example
(4). Stressed syllables are marked as o , with no
distinction between primary and secondary stress,
since both degrees of stress are equally relevant for
rhythmic alternation.
Syllabic trochees: left-to-right:
(3) (o
(o
(o
(o
(o
o)
o)
o)
o)
o)
(o )
(o o)
(o o) (o )
(o o) (o o)
Syllabic trochees: right-to-left:

(4)
(o
o
(o
o
(o
o)
(o
o)
(o
o)
o)
(o o)
o) (o o)
(o o) (o o)
In Example (3), based on languages such as Icelandic,

Hungarian, or Maranungku (Hayes, 1995), rhythmic
alternation proceeds from left to right. Note that in
words with an odd number of syllables, the final
group is not binary, but for this very reason, stress
alternation is regular. Systems minimally departing
from Example (3) (e.g., Pintupi; Hayes, 1995) lack
stress on the final syllable. As a result, words with an
odd number of syllables have adjacent weak beats, as
in (o o) o and (o o) (o o) o, with a disruption of the
alternating pattern at the right edge. In Example (4),
based on languages such as Warao or Cavinen a
(Hayes, 1995), rhythmic alternation proceeds from
right to left. Systems minimally departing from Example (4) are those in which word-initial syllables
attract stress, yielding adjacent weak beats in words
with an odd number of syllables, as in (o o) o (o o).
This is the so-called initial dactyl effect (described in
Prince (1983), Cohn (1993), and Hayes (1995),
among others).
Alternation in stress systems that take the mora as
the basic beat is of a less robust nature. In this case,
the alternating pattern is computed over moras, yet
stress is realized on syllables. Crucially, regular alternation at the level of moraic groupings need not be
realized as regular alternation at the level of syllables.
Groupings into moraic trochees include two light
syllables (o L oL), or one heavy syllable (o H). However,
whereas forms with all light syllables will exhibit
perfect stress alternation, forms that combine light
and heavy syllables may not, as in (o H) (o L oL) (o H),
wherein stresses on the first two syllables clash (e.g.,
Fijian; Hayes, 1995). Likewise, groupings into
moraic iambs include two light syllables (o L oL), one
heavy syllable (o H), or a light followed by a heavy (oL
o H). In this case, again, a stress clash may arise at the
level of the syllables, as in (o`H) (o`L oL) (o H), even
though alternation at the level of the mora is perfectly
regular (e.g., Asheninca; Hayes, 1995).
The Trochaic/Iambic Law
Alternating patterns are well attested for three out of

four grouping types for all, that is, other than the
system of syllabic iambs. Only a single case of what
appear to be syllabic iambs (Weri) is listed in Hayes
(1981). Hayes (1995) eliminated syllabic iambs from
the set of possible groupings, thus introducing an
asymmetry into the system and motivating it by the
general cognitive properties of rhythmic alternation.
According to Woodrow (1951: 1233), with equal
temporal spacing, a regularly recurring, relatively
greater intensity exerts a group-beginning effect, and
a regularly recurring, relatively greater duration a
group-ending effect. In other words, alternation in
intensity produces trochaic grouping, and alternation
in duration produces iambic grouping. These asymmetric grouping tendencies are evidenced in music
(Cooper and Meyer, 1960: 10) and, according to
Hayes (1995: 7982) and Prince (1990), are also
evidenced in rhythmic patterns in language.
Groupings in the two moraic systems exhibit a crucial asymmetry of the predicted sort: trochaic groupings, (o L oL), (o H), are even, and iambic groupings, (oL
o L), (o H), (oL o H), may be uneven (Hayes, 1995;
Kager, 1995; Prince, 1990). There is no (oH oL) trochaic grouping; in fact, systems of moraic trochees
exhibit so-called trochaic shortening, as in Fijian
(Hayes, 1995), reinforcing the strong preference for
even trochees. Moreover, (oL oH) appears to be the
preferred iamb, as evidenced by numerous cases
of iambic lengthening, the effect of which is to eliminate feet of the (oL o L) shape and reaffirm the preference for uneven iambs (Hayes, 1995). However, in the
syllabic systems, all syllables are treated as equals, and
as a result, all groups are even. Because of the preference for uneven iambic groupings, syllabic systems
are inconsistent with iambic grouping. The strong
theoretical position is that syllabic iambs are either
highly disfavored (Prince, 1990) or are eliminated
from the system (Hayes, 1995). It should be noted,
however, that Gordons (2002: 522) broad survey of
syllabic stress systems included eight cases of syllabic
iambic groupings, thus strengthening the empirical
basis for positing this type of stress alternation.
Other Alternating Patterns in Word Phonology
Though most directly, and predominantly, manifested

in stress systems, rhythmic alternation is not entirely
restricted to them. Equally relevant, though not as well
attested, are rhythmic alternations in phenomena other
than stress, such as vowel length or vowel quality,
which ultimately belong to the phonological devices
for expressing degrees of prominence. Interestingly,
clear cases of this type are characterized by moraic
groupings. Three such cases are presented here. The
Creek language exhibits covert rhythmic alternation:
the place of tonal accent is computed by virtue of
left-to-right iambic grouping, yet only the rightmost
iambic group is phonologically marked by tone
(Hayes, 1995). Next, Chickasaw exhibits rhythmic
lengthening that in no obvious way relates to stress,
and affects every second light syllable computed
from left to right (Munro and Willmond, 1994;
Hayes, 1995). Finally, Old Church Slavonic has an
alternating pattern defined by Havlks law, exhibited
by high lax vowels (Isac enko, 1970). When two open
syllables, both containing high lax vowels, are adjacent, the first one is lowered to a mid-vowel. Moreover,
grouping proceeds from right to left: in words with
only high lax vowels, lowering takes effect in every
622 Rhythmic Alternations
even-numbered syllable from the end. In sum, cases of

rhythmic alternation unrelated to stress provide strong
evidence for full integration of rhythmic alternation
into the inventory of possible phonological patterns.
Rhythmic Alternation in Phrasal Domains

Though the domain of rhythmic alternation is wordsized for only a subset of languages characterized by
stress, any language with word stress could potentially
exhibit rhythmic distribution of stresses within phrasal domains. Moreover, though rhythmic alternation
at the word level is subject to a tight phonological
organization, inter-stress intervals in the phrasal domain are under looser phonological control. This is in
part because regular distribution of stress in phrasal
domains crucially depends on word size. However,
though there is no evidence for clearly defined groupings comparable to those at the word level, phonological regularizations of phrasal rhythm do occur, as in
the well-studied case of the rhythm rule in English
(Liberman and Prince, 1977; Hayes, 1984, 1995;
Selkirk, 1984; Kager and Visch, 1988; and references
therein). As illustrated in Example (5), stress in thirte en shifts leftward when followed by a stronger
stress, resulting in an alternating rhythmic pattern.
Rhythmic alternation is represented here with grids,
first introduced by Liberman and Prince (1977) to
accommodate this type of rhythmic adjustment:
(5)
x
x
x
x
x
x
x x
x
x x
x
thirteen men!thirteen men
Crucially, stress shifts leftward only if there is an appropriate landing site: stress will only shift onto a syllable
with secondary prominence. However, rhythmic adjustments of this type are not mandatory, but, rather, point
at a strong tendency toward eurhythmy that is, toward
rhythmic alternation in phrasal domains. Comparable
rhythmic effects have been observed in German
(Kiparsky, 1966) and Dutch (Kager and Visch, 1988),
as well as in Polish, Italian, and Greek (Nespor and
Vogel, 1989) (for additional cases, see Hayes (1995)).
The Phonetic Realization of Rhythm

In studies of speech rhythm, a well-known proposal
is the division of languages into two rhythmic
types: stress timed and syllable timed (Pike, 1945;
Abercrombie, 1967). In the former type, stressed syllables occur at equal time intervals, and in the latter
type, all syllables occur in such fashion. Thus, the
former type is characterized by stress-based isochrony,
and the latter by syllable-based isochrony. Note that
the assumed isochronous intervals provide a basis for

an alternating rhythm only in stress-timed languages.
Phonetic studies have refuted this sharp division.
Dauer (1983) failed to identify differences in the duration of inter-stress intervals between Spanish, classified as a syllable-timed language, and English, a
stress-timed language. Moreover, Wenk and Wioland
(1982) demonstrated that French, presumably a syllable-timed language, is not characterized by an increase of word duration proportional to the number
of syllables. However, though there is no clear case
for isochrony at the level of production, it has been
found that there is a strong tendency toward isochrony at the perceptual level: inter-stress intervals
tend to be perceived as more isochronous than what
would be expected from their actual durations
(Lehiste, 1977; Scott and Isard, 1985). A further important finding is that perception of the isochrony of
inter-stress intervals is stronger within, than across,
linguistic constituents (Donovan and Darwin, 1979).
In sum, phonetic studies of rhythm have identified
inter-stress intervals as intervals that tend to be
perceived as such, though not necessarily recurring
with relatively stable durations. Yet, this rhythmic
alternation is evidenced only in a fairly weak sense.
See also: Foot; Meter; Metrical Phonology; Phrasal Stress;
Poetry: Stylistic Aspects; Rhyme; Rhythm; Semantics of

Prosody; Word Stress.
Bibliography
Chicago: Aldine Publ. Co.
Cohn A C (1993). The initial dactyl effect in Indonesian.
Linguistic Inquiry 24, 372381.
Cooper G W & Meyer L B (1960). The rhythmic structure
of music. Chicago: Chicago University Press.
Dauer R (1983). Stress-timing and syllable-timing reanalyzed. Journal of Phonetics 11, 5262.
Donovan A & Darwin C J (1979). The perceived rhythm of
speech. In Proceedings of the Ninth International Congress of Phonetic Sciences, 611 August, 1979, Copenhagen, vol. 2. Copenhagen: Institute of Phonetics.
Elenbaas N & Kager R (1999). Ternary rhythm and the
lapse constraint. Phonology 16, 273329.
Fraisse P (1978). Time and rhythm perception. In
Carterette E C & Friedman M P (eds.) Handbook of
perception, vol. VIII: Perceptual coding. New York:
Academic Press. 203254.
Gordon M (2002). A factorial typology of quantity-insensitive stress. Natural Language and Linguistic Theory
20, 491552.
Halle M & Idsardi W (1995). General properties of stress
and metrical structure. In Goldsmith J (ed.) The handbook of phonological theory. Blackwell Publ. 403443.
Halle M & Vergnaud J-R (1987). An essay on stress. Cambridge: The MIT Press.
Riau Indonesian 623

Hayes B (1981). A metrical theory of stress rules. Ph.D. diss.,
Massachusetts Institute of Technology Cambridge, MA.
Hayes B (1984). The phonology of rhythm in English.
studies. Chicago: The University of Chicago Press.
Isac enko A V (1970). East Slavic morphophonemics and
the treatment of the jers in Russian: a revision of Havlks
law. International Journal of Slavic Linguistics and
Poetics 13, 73124.
Kager R (1995). The metrical theory of word stress. In
Goldsmith J (ed.) The handbook of phonological theory.
Blackwell Publ. 367402.
Kager R & Visch E (1988). Metrical constituency and
rhythmic adjustment. Phonology 5, 2171.
Kiparsky P (1966). Ueber den Deutschen Akzent. Studia
Grammatica 7, 6998.
Lehiste I (1977). Isochrony reconsidered. Journal of Phonetics 5, 253263.
Lerdahl F & Jackendoff R (1983). A generative theory of
tonal music. Cambridge, MA: The MIT Press.
Levin J (1988). Generating ternary feet. Texas Linguistic
Forum 29, 97113.
Liberman M (1975). The intonational system of English.
Ph.D. diss., Massachusetts Institute of Technology
Cambridge, MA.
Liberman M & Prince A S (1977). On stress and linguistic

rhythm. Linguistic Inquiry 8, 249336.
Munro P & Willmond C (1994). Chickasaw. An analytical
dictionary. Norman, OK and London: University of
Oklahoma Press.
Phonology 6, 69116.
Pike K L (1945). The intonation of American English. Ann
Arbor, MI: University of Michigan Press.
Prince A S (1983). Relating to the grid. Linguistic Inquiry
14, 19100.
Prince A S (1990). Quantitative consequences of rhythmic organization. Parasession on the syllable in phonetics and phonology. Chicago Linguistic Society 26,
355398.
Scott D R & Isard S D (1985). Perceptual isochrony in
English and French. Journal of Phonetics 13, 155162.
Selkirk E O (1984). Phonology and syntax: the relation
between sound and structure. Cambridge, MA: The
MIT Press.
Wenk B J & Wioland F (1982). Is French really syllabletimed? Journal of Phonetics 10, 193216.
Woodrow H (1951). Time perception. In Stevens S S (ed.)
Handbook of experimental psychology. New York:
Wiley. 12341236.
Riau Indonesian
D Gil, Max Planck Institute for Evolutionary
Anthropology, Leipzig, Germany
Riau Indonesian is the variety of colloquial Indonesian

spoken by the inhabitants of the Indonesian province
of Riau, which encompasses parts of east-central
Sumatra plus a large number of adjacent smaller
islands. Riau Indonesian is one of many distinct local
varieties of colloquial Indonesian spoken throughout the archipelago, such as, for example, Jakarta
Indonesian. The population of Riau province, numbering close to 5 million people, is linguistically and
ethnically heterogeneous. Although the indigenous
population is mostly Malay, a majority of the presentday inhabitants are migrants from other provinces,
speaking a variety of other languages. Riau Indonesian is acquired as a native language by most or
all children growing up in Riau province, whatever
their ethnicity. It is the language most commonly used
as a lingua franca for interethnic communication,
and, in addition, it is gradually replacing other
languages and dialects as a vehicle for intraethnic
communication.
Riau Indonesian is quite different from Standard

Indonesian, a language familiar to many general linguists from a substantial descriptive and theoretical
literature. Riau Indonesian is also distinct from a
set of dialects generally referred to as Riau Malay,
used in Riau province by ethnic Malays, primarily
for intraethnic communication. In addition, Riau
Indonesian is distinguished from another set of
Malayic dialects spoken by various indigenous
peoples in Riau province (Orang Asli, Orang Sakai,
Orang Akit, Orang Hutan, and Orang Laut). Finally,
Riau Indonesian is also different from the variety
of Malay/Indonesian sometimes referred to as Bazaar
Malay, which is used by the ethnic Chinese residents
of Riau province when speaking to non-Chinese,
and by the non-Chinese when speaking to the
ethnic Chinese. Thus, the sociolinguistic situation in
Riau province is one of great complexity: speakers
of Riau Indonesian are often fluent in several other
varieties of Malay/Indonesian, as well as in other languages, such as Minangkabau and Javanese.
From a general typological perspective, Riau Indonesian is a strongly isolating language, with no
inflectional morphology and relatively little derivational morphology or compounding. It is also a
Riau Indonesian 623

Hayes B (1981). A metrical theory of stress rules. Ph.D. diss.,
Massachusetts Institute of Technology Cambridge, MA.
Hayes B (1984). The phonology of rhythm in English.
studies. Chicago: The University of Chicago Press.
Isacenko A V (1970). East Slavic morphophonemics and
the treatment of the jers in Russian: a revision of Havlks
law. International Journal of Slavic Linguistics and
Poetics 13, 73124.
Kager R (1995). The metrical theory of word stress. In
Goldsmith J (ed.) The handbook of phonological theory.
Blackwell Publ. 367402.
Kager R & Visch E (1988). Metrical constituency and
rhythmic adjustment. Phonology 5, 2171.
Kiparsky P (1966). Ueber den Deutschen Akzent. Studia
Grammatica 7, 6998.
Lehiste I (1977). Isochrony reconsidered. Journal of Phonetics 5, 253263.
Lerdahl F & Jackendoff R (1983). A generative theory of
tonal music. Cambridge, MA: The MIT Press.
Levin J (1988). Generating ternary feet. Texas Linguistic
Forum 29, 97113.
Liberman M (1975). The intonational system of English.
Ph.D. diss., Massachusetts Institute of Technology
Cambridge, MA.
Liberman M & Prince A S (1977). On stress and linguistic

rhythm. Linguistic Inquiry 8, 249336.
Munro P & Willmond C (1994). Chickasaw. An analytical
dictionary. Norman, OK and London: University of
Oklahoma Press.
Phonology 6, 69116.
Pike K L (1945). The intonation of American English. Ann
Arbor, MI: University of Michigan Press.
Prince A S (1983). Relating to the grid. Linguistic Inquiry
14, 19100.
Prince A S (1990). Quantitative consequences of rhythmic organization. Parasession on the syllable in phonetics and phonology. Chicago Linguistic Society 26,
355398.
Scott D R & Isard S D (1985). Perceptual isochrony in
English and French. Journal of Phonetics 13, 155162.
Selkirk E O (1984). Phonology and syntax: the relation
between sound and structure. Cambridge, MA: The
MIT Press.
Wenk B J & Wioland F (1982). Is French really syllabletimed? Journal of Phonetics 10, 193216.
Woodrow H (1951). Time perception. In Stevens S S (ed.)
Handbook of experimental psychology. New York:
Wiley. 12341236.
Riau Indonesian
D Gil, Max Planck Institute for Evolutionary
Anthropology, Leipzig, Germany
Riau Indonesian is the variety of colloquial Indonesian

spoken by the inhabitants of the Indonesian province
of Riau, which encompasses parts of east-central
Sumatra plus a large number of adjacent smaller
islands. Riau Indonesian is one of many distinct local
varieties of colloquial Indonesian spoken throughout the archipelago, such as, for example, Jakarta
Indonesian. The population of Riau province, numbering close to 5 million people, is linguistically and
ethnically heterogeneous. Although the indigenous
population is mostly Malay, a majority of the presentday inhabitants are migrants from other provinces,
speaking a variety of other languages. Riau Indonesian is acquired as a native language by most or
all children growing up in Riau province, whatever
their ethnicity. It is the language most commonly used
as a lingua franca for interethnic communication,
and, in addition, it is gradually replacing other
languages and dialects as a vehicle for intraethnic
communication.
Riau Indonesian is quite different from Standard

Indonesian, a language familiar to many general linguists from a substantial descriptive and theoretical
literature. Riau Indonesian is also distinct from a
set of dialects generally referred to as Riau Malay,
used in Riau province by ethnic Malays, primarily
for intraethnic communication. In addition, Riau
Indonesian is distinguished from another set of
Malayic dialects spoken by various indigenous
peoples in Riau province (Orang Asli, Orang Sakai,
Orang Akit, Orang Hutan, and Orang Laut). Finally,
Riau Indonesian is also different from the variety
of Malay/Indonesian sometimes referred to as Bazaar
Malay, which is used by the ethnic Chinese residents
of Riau province when speaking to non-Chinese,
and by the non-Chinese when speaking to the
ethnic Chinese. Thus, the sociolinguistic situation in
Riau province is one of great complexity: speakers
of Riau Indonesian are often fluent in several other
varieties of Malay/Indonesian, as well as in other languages, such as Minangkabau and Javanese.
From a general typological perspective, Riau Indonesian is a strongly isolating language, with no
inflectional morphology and relatively little derivational morphology or compounding. It is also a
624 Riau Indonesian
language with very flexible word order. Perhaps

the most striking feature of Riau Indonesian is the
pervasiveness of underspecification, i.e., the absence
of obligatory overt grammatical expression for a
wide variety of semantic categories, including number, definiteness, tense, aspect, thematic role, and
ontological type. On the basis of these characteristics,
Riau Indonesian has been argued to have a simple
grammar, lacking much of the machinery central to
most grammatical theories. Syntactically, it is said
to have a single open syntactic category, that is to
say, no distinction between nouns, verbs, adjectives,
and prepositions, or between lexical categories and
phrasal ones. Semantically, it is claimed that when
two or more expressions are combined, the meaning of the combination is usually associated with
the meanings of the constituents in a vague and

underspecified fashion.
See also: Austronesian Languages: Overview; Indonesia:
Language Situation; Malay; Primitive Languages; Word
Classes/Parts of Speech: Overview.
Bibliography
Gil D (1994). The structure of Riau Indonesian. Nordic
Journal of Linguistics 17, 179200.
Gil D (2001). Creoles, complexity and Riau Indonesian.
Linguistic Typology 5, 325371.
Gil D (2005). Isolating-monocategorial-associational language. In Cohen H & Lefebvre C (eds.) Categorization
in cognitive science. Oxford: Elsevier.
Richards, Ivor Armstrong (18931979)

N Kerecuk, London, UK
Ivor Armstrong Richards, a British literary theorist,

critic, poet and educator, was born in Cheshire,
United Kingdom, in 1893. He was educated at Clifton
College in Bristol. Originally, he read history but
changed to moral sciences at Magdalene College graduating in 1915. After recovering from tuberculosis, he
decided to study medicine but was invited to teach
at the new English School (faculty) at Magdalene
College in 1919, where he stayed until 1939. He
traveled and taught throughout the world, especially
in China and Japan. In 1944, he became a professor in
the Department of Education at Harvard University,
where he taught until 1963. He maintained his
contacts with Magdalene College and returned to
Cambridge in 1974. He died in Cambridge in 1979.
In 1918, he met Charles Kay Ogden (18891952),
beginning a cooperation in philosophical linguistics.
In 1957, Ogden wrote about the circumstances in
which they met and how they discussed some of the
controversies that were debated in Mind and particularly on how quickly Richards and Ogden produced
an outline for The meaning of meaning.
In 1922, The foundations of aesthetics was published by Richards, Ogden, and James Wood (painter). In this work, there is a discussion of beauty that
was treated more comprehensively in The meaning of
meaning: it contains 16 senses of the word beauty
examined psychologistically. Ogden was a student of
and was influenced by the heritage of the thoughts of
Lady Victoria Welby (Gregory) (18371912), an independent scholar. In the late 19th century, she published
papers on meaning, significs, in Mind and The
Monist. Lady Welby started the signific movement
especially through the Signifische Kring in The
Netherlands in 1917, when she met the Dutch poet
and psychotherapist Frederik van Eeden (1860
1932), and she would later announce a Welby prize
for original ideas on meaning that led to the development of Basic English. She was also active in the Vienna
Circle and, of necessity, was influenced by the ideas that
migrated to Vienna. Lady Welby also corresponded
with C. S. Peirce (18391914). Lady Welbys thoughts
influenced both Richards and Ogden.
In 1923, Ogden and Richards published The meaning of meaning. A study of the influence of language
upon thought and of the science of symbolism. This
work made the authors very famous. For example,
Bertrand Russell and Edward Sapir discussed and
adopted this book as standard university reading
text. In this work, they investigated the role that
words play on thought, the symbolic and emotional
functions of meaning, and they sought to improve the
incipient behaviorism. Their semiotic triangle
thought or reference, symbol, and referent has
been disseminated among the various language
sciences. The manner by which they described the
relation between symbol and referent has been criticized within the discussions of theory of meaning.
One of the perceived failures is that their model fails
to account for the cognitive component of human
communication. Richards also argued that human
thought is metaphoric in this work.
624 Riau Indonesian
language with very flexible word order. Perhaps

the most striking feature of Riau Indonesian is the
pervasiveness of underspecification, i.e., the absence
of obligatory overt grammatical expression for a
wide variety of semantic categories, including number, definiteness, tense, aspect, thematic role, and
ontological type. On the basis of these characteristics,
Riau Indonesian has been argued to have a simple
grammar, lacking much of the machinery central to
most grammatical theories. Syntactically, it is said
to have a single open syntactic category, that is to
say, no distinction between nouns, verbs, adjectives,
and prepositions, or between lexical categories and
phrasal ones. Semantically, it is claimed that when
two or more expressions are combined, the meaning of the combination is usually associated with
the meanings of the constituents in a vague and

underspecified fashion.
See also: Austronesian Languages: Overview; Indonesia:
Language Situation; Malay; Primitive Languages; Word
Classes/Parts of Speech: Overview.
Bibliography
Gil D (1994). The structure of Riau Indonesian. Nordic
Journal of Linguistics 17, 179200.
Gil D (2001). Creoles, complexity and Riau Indonesian.
Linguistic Typology 5, 325371.
Gil D (2005). Isolating-monocategorial-associational language. In Cohen H & Lefebvre C (eds.) Categorization
in cognitive science. Oxford: Elsevier.
Richards, Ivor Armstrong (18931979)

N Kerecuk, London, UK
Ivor Armstrong Richards, a British literary theorist,

critic, poet and educator, was born in Cheshire,
United Kingdom, in 1893. He was educated at Clifton
College in Bristol. Originally, he read history but
changed to moral sciences at Magdalene College graduating in 1915. After recovering from tuberculosis, he
decided to study medicine but was invited to teach
at the new English School (faculty) at Magdalene
College in 1919, where he stayed until 1939. He
traveled and taught throughout the world, especially
in China and Japan. In 1944, he became a professor in
the Department of Education at Harvard University,
where he taught until 1963. He maintained his
contacts with Magdalene College and returned to
Cambridge in 1974. He died in Cambridge in 1979.
In 1918, he met Charles Kay Ogden (18891952),
beginning a cooperation in philosophical linguistics.
In 1957, Ogden wrote about the circumstances in
which they met and how they discussed some of the
controversies that were debated in Mind and particularly on how quickly Richards and Ogden produced
an outline for The meaning of meaning.
In 1922, The foundations of aesthetics was published by Richards, Ogden, and James Wood (painter). In this work, there is a discussion of beauty that
was treated more comprehensively in The meaning of
meaning: it contains 16 senses of the word beauty
examined psychologistically. Ogden was a student of
and was influenced by the heritage of the thoughts of
Lady Victoria Welby (Gregory) (18371912), an independent scholar. In the late 19th century, she published
papers on meaning, significs, in Mind and The
Monist. Lady Welby started the signific movement
especially through the Signifische Kring in The
Netherlands in 1917, when she met the Dutch poet
and psychotherapist Frederik van Eeden (1860
1932), and she would later announce a Welby prize
for original ideas on meaning that led to the development of Basic English. She was also active in the Vienna
Circle and, of necessity, was influenced by the ideas that
migrated to Vienna. Lady Welby also corresponded
with C. S. Peirce (18391914). Lady Welbys thoughts
influenced both Richards and Ogden.
In 1923, Ogden and Richards published The meaning of meaning. A study of the influence of language
upon thought and of the science of symbolism. This
work made the authors very famous. For example,
Bertrand Russell and Edward Sapir discussed and
adopted this book as standard university reading
text. In this work, they investigated the role that
words play on thought, the symbolic and emotional
functions of meaning, and they sought to improve the
incipient behaviorism. Their semiotic triangle
thought or reference, symbol, and referent has
been disseminated among the various language
sciences. The manner by which they described the
relation between symbol and referent has been criticized within the discussions of theory of meaning.
One of the perceived failures is that their model fails
to account for the cognitive component of human
communication. Richards also argued that human
thought is metaphoric in this work.
Richards, Ivor Armstrong (18931979) 625
Richards regarded himself as a philosopher with

interests in psychology, and his earliest papers were
on aesthetics published in Athenaeum in 1919: Art
and science, Emotion and art, Four fermented aesthetics, and The instruments of criticism: expression. At Magdalene College, he lectured on literary
criticism (novel) and theory of criticism. He developed courses on practical criticism using his method
of close reading and courses on Samuel T. Coleridge,
rhetorical theory, and British moralists.
He stated that poetry, which often stood for literature or art in his arguments, was central to everything, and soon he became a theorist of the arts. He
was very interested in the effect that reading had on
the minds of readers: He argued that each literary
work read modified the person who read it. In
1925, he published Principles of literary criticism,
which he had started writing in 1923. In this work,
he aimed to place the arts at the forefront of all
values because he regarded arts as the supreme
mode of communication, as he wrote in a letter to
D. E. Pilley (November 19, 1923). He intended to
demonstrate the value of poetry within the context
of the contemporary psychological framework. The
norms and practices of science hindered the ability of
man to read and write poetry, according to Richards,
who aimed to reinstate poetic expression or discourse. In his 1926 Science and poetry (revised in
1935), he introduced pseudo-statements (misinterpreted then as false statements) as poetic utterances
that do not need to be judged for truth conditions as
does a scientific statement. Because he was interested
in improving the reading skills of the university students, he dedicated much effort to this end, which
resulted in the publication of Philosophy of rhetoric
in 1936 and Interpretation in teaching in 1938. He
stated that he rejected Aristotelian and other traditional rhetoricians thought and argued that rhetoric
should be a philosophical inquiry into how words
work in discourse. He applied these principles to
Mencius, the Chinese philosopher, while he was
traveling in China. He learned some Chinese, developed a simplified form of literal translations, and his
study was published in 1932 as Mencius on the mind:
experiments in multiple definition.
Richards moved away from literary discussion after
publishing Practical criticism and turned to educational theory and language teaching. By 1931, he
returned to the ideas of a system of simplified language, Basic English, with Ogden. This led him to
produce a list of 850 words with which almost all
communicative needs can be expressed. He dedicated
the rest of his life to this enterprise. He promoted
world literacy. In his cooperation with Christine
M. Gibson, he pioneered the use of audiovisual media
resources to the teaching of English as a foreign language. His works represent an important chapter
in the history of language teaching methodology in
general and specifically in English.
See also: Literary Theory and Stylistics; Ogden, Charles
Kay (18891957); Peirce, Charles Sanders (18391914);
Russell, Bertrand (18721970); Sapir, Edward (18841939).
Bibliography
Eagleton T (ed.) (2001). I. A. Richards: selected works
19191938 (10 vols). London: Routledge.
Foss S K et al. (1991). Contemporary perspectives on rhetoric (2nd edn.). Prospect Heights, IL: Waveland.
Martins W (2001). A palavra escrita. Histo ria do livro, da
imprensa e da biblioteca. Sa o Paulo, Brazil: Editora
A tica.
Noordegraaf J (1991). Communication, significs and linguistics. Beitrage zur Geschichte der Sprachwissenshaft
1, 6973.
Ogden C K (1994). C. K. Ogden and linguistics. Gordon T
(ed.) (5 vols). London: Routledge/Thoemmes.
Ogden C K & Richards I A (1923). The meaning of meaning. A study of the influence of language upon thought
and the science of symbolism. London: Routledge &
Keagan Paul.
Peirce C S (1977). Semiotics and significs: the correspondence between Charles S. Peirce and Victoria Lady Welby
(edited by Hardwick C S). Bloomington: Indiana University Press.
Richards I A (1936). The philosophy of rhetoric. New York:
Oxford University Press.
Richards I A (1965). Principles of literary criticism. London: Keagan Paul.
Schmitz H W (1985). Verstandigungshandlungen eine wissenschaftshistorische Rekonstruktion der Anfange der signifischen Bewegung in den Niederlanden (18921926).
Habitilationschrift Friedrich-Wlihelms-Universitat, Bonn.
Steiner G (1975). After Babel. Aspects of language and
translation. Oxford: Oxford University Press.
Verburg P A (1998). Language and its functions. A historicocritical study of views concerning the functions of language from the pre-humanist philology of Orleans to the
rationalist philology of Bopp (Salmon P, trans., in consultation with Klijnsmit A J). Amsterdam: Benjamins.
Welby Victoria Lady (1893). Meaning and metaphor. The
Monist 3(4), 510525 (Reprinted 1911).
Welby Victoria Lady (1896). Sense, meaning and interpretation. Mind, 5:17 (1896), 2437 and Mind, 5:18
(1896), 186202. In Welby (1985).
Welby Victoria Lady (1897). Grains of sense. London:
Dent.
Welby Victoria Lady (1903). What is meaning. Studies in
the development of significance. London: Macmillan.
(Reprinted in 1983 by Benjamins, Amsterdam.)
Welby Victoria Lady (1911). Significs. In The encyclopaedia Britannica. A dictionary of arts, sciences, literature
and general information (11th edn., vol. 25). Cambridge,
626 Richards, Ivor Armstrong (18931979)

UK: Cambridge University Press. 7881. (Reprinted in
Peirce 1977.)
Welby Victoria Lady (1985). Significs and Language: The
Articulate Form of Our Expressive and Interpretive
Resources. Reprinted original edition London, 1911.

Edited by H. Walter Schmitz. Amsterdam and Philadelphia: John Benjamins Publishing Company.
Richter, Elise (18651942)

C Neis, Potsdam University, Potsdam, Germany
Elise Richter was born March 2, 1865 in Vienna and

died June 21, 1943 in Theresienstadt (now Terezn,
Czech Republic). Richter was one of the leading
figures of Romance philology and phonetics in the
20th century. Having enjoyed the benefits of private
education, she was the first woman in Austria who
passed the matura and one of the first three women
who were allowed to study at the University of
Vienna. Richter was the first female student at the
department of Romance philology, where she studied
with Adolf Mussafia (18351905) and Wilhelm
Meyer-Lubke (18611936). She was awarded a doctorate in 1901. Her dissertation Zur Entwicklung
der romanischen Wortstellung aus der lateinischen
(1903) was inspired by Meyer-Lubkes Grammatik
der romanischen Sprachen (18901902). It was also
Meyer-Lubke who encouraged Richter to apply for
a Privatdozentur. In 1903 she presented her habilitation Ab im Romanischen to the Philosophical Faculty
at Vienna. Richter tried to demonstrate that the Latin
preposition ab is still present in several prepositions in
Romance languages (Italian: da, French: a`, Spanish:
a etc.). In 1907, the Ministry confirmed the venia
legendi and appointed Richter as the first female Privatdozent in Austria and Germany. Being a woman,
she never became ordinaria, but due to her generally
acknowledged competence she became extraordinaria in 1921. As a result of the annexation of Austria
by Germany in 1938, Richter was forced to resign
because of her Jewish origin. In 1939, the International Federation of University Women offered some
financial support to Elise and her sister Helene,
had they agreed to immigrate to England. But Elise
Richter refused to leave her home country (Alte
Ba ume verpflanzt man nicht). In 1942, she was
deported to the concentration camp Theresienstadt,
where she died on June 21, 1943.
At an early stage in her writings, Richter developed
an independent method of research and exceeded her
mentor Meyer-Lubke in originality. Her article Die
Rolle der Semantik in der historischen Grammatik
(1909) showed her independent manner of thinking
as she applied the Neogrammarians achievements

(especially the analogy) to the field of semantics,
which until that time had been considered as a domain separate from grammar. Richters research was
particularly focused on the historical development of
Romance languages. In Der innere Zusammenhang in
der Entwicklung der romanischen Sprachen (1911),
Richter explained the development of Romance languages from Latin as a result of a change in wordorder. In her attempt to incorporate psychological
aspects into her research on syntax and semantics,
she adopted the conceptions of Wilhelm Wundt and
Karl Vossler, but she never succumbed entirely to
Vosslers idealism.
Richter also made important contributions to phonetics (a domain included in her teaching duties),
particularly to experimental phonetics. Stimulated
by her Vienna colleague Trubetzkoy, Richter became
a pioneer of phonology in Romance languages.
In many of her writings, Richter embedded her research within a larger context, including general linguistics, German linguistics, Romance literature, etc.
Most of Elise Richters innovative work has had a
remarkable influence on the subsequent development
of linguistics.
See also: Meyer-Lubke, Wilhelm (18611936); Trubetskoy,
Nikolai Sergeievich, Prince (18901938); Vossler, Karl

(18721949); Wundt, Wilhelm (18321920).
Bibliography
Christmann H H (1980). Frau und Ju din an der Universita t. Die Romanistin Elise Richter (Wien 1865 Theresienstadt 1943). Mainz (Abhandlungen der Geistes- und
Sozialwissenschaftlichen Klasse 1980, 2) & Wiesbaden:
Franz Steiner Verlag.
Christmann H H & Hausmann F-R (eds.) (1989). In
Verbindung mit Manfred Briegel (1989). Deutsche
und o sterreichische Romanisten als Verfolgte des Nationalsozialismus. Mit einer bio-bibliographischen Dokumentation der verfolgten Romanisten (Romanica et
Comparatistica, vol. 10). Tubingen: Stauffenburg.
Meyer-Lubke W (18901902). Grammatik der romanischen Sprachen (4 vols). Leipzig: Fues/Reisland.
626 Richards, Ivor Armstrong (18931979)

UK: Cambridge University Press. 7881. (Reprinted in
Peirce 1977.)
Welby Victoria Lady (1985). Significs and Language: The
Articulate Form of Our Expressive and Interpretive
Resources. Reprinted original edition London, 1911.

Edited by H. Walter Schmitz. Amsterdam and Philadelphia: John Benjamins Publishing Company.
Richter, Elise (18651942)

C Neis, Potsdam University, Potsdam, Germany
Elise Richter was born March 2, 1865 in Vienna and

died June 21, 1943 in Theresienstadt (now Terezn,
Czech Republic). Richter was one of the leading
figures of Romance philology and phonetics in the
20th century. Having enjoyed the benefits of private
education, she was the first woman in Austria who
passed the matura and one of the first three women
who were allowed to study at the University of
Vienna. Richter was the first female student at the
department of Romance philology, where she studied
with Adolf Mussafia (18351905) and Wilhelm
Meyer-Lubke (18611936). She was awarded a doctorate in 1901. Her dissertation Zur Entwicklung
der romanischen Wortstellung aus der lateinischen
(1903) was inspired by Meyer-Lubkes Grammatik
der romanischen Sprachen (18901902). It was also
Meyer-Lubke who encouraged Richter to apply for
a Privatdozentur. In 1903 she presented her habilitation Ab im Romanischen to the Philosophical Faculty
at Vienna. Richter tried to demonstrate that the Latin
preposition ab is still present in several prepositions in
Romance languages (Italian: da, French: a`, Spanish:
a etc.). In 1907, the Ministry confirmed the venia
legendi and appointed Richter as the first female Privatdozent in Austria and Germany. Being a woman,
she never became ordinaria, but due to her generally
acknowledged competence she became extraordinaria in 1921. As a result of the annexation of Austria
by Germany in 1938, Richter was forced to resign
because of her Jewish origin. In 1939, the International Federation of University Women offered some
financial support to Elise and her sister Helene,
had they agreed to immigrate to England. But Elise
Richter refused to leave her home country (Alte
Baume verpflanzt man nicht). In 1942, she was
deported to the concentration camp Theresienstadt,
where she died on June 21, 1943.
At an early stage in her writings, Richter developed
an independent method of research and exceeded her
mentor Meyer-Lubke in originality. Her article Die
Rolle der Semantik in der historischen Grammatik
(1909) showed her independent manner of thinking
as she applied the Neogrammarians achievements

(especially the analogy) to the field of semantics,
which until that time had been considered as a domain separate from grammar. Richters research was
particularly focused on the historical development of
Romance languages. In Der innere Zusammenhang in
der Entwicklung der romanischen Sprachen (1911),
Richter explained the development of Romance languages from Latin as a result of a change in wordorder. In her attempt to incorporate psychological
aspects into her research on syntax and semantics,
she adopted the conceptions of Wilhelm Wundt and
Karl Vossler, but she never succumbed entirely to
Vosslers idealism.
Richter also made important contributions to phonetics (a domain included in her teaching duties),
particularly to experimental phonetics. Stimulated
by her Vienna colleague Trubetzkoy, Richter became
a pioneer of phonology in Romance languages.
In many of her writings, Richter embedded her research within a larger context, including general linguistics, German linguistics, Romance literature, etc.
Most of Elise Richters innovative work has had a
remarkable influence on the subsequent development
of linguistics.
See also: Meyer-Lubke, Wilhelm (18611936); Trubetskoy,
Nikolai Sergeievich, Prince (18901938); Vossler, Karl

(18721949); Wundt, Wilhelm (18321920).
Bibliography
Christmann H H (1980). Frau und Judin an der Universitat. Die Romanistin Elise Richter (Wien 1865 Theresienstadt 1943). Mainz (Abhandlungen der Geistes- und
Sozialwissenschaftlichen Klasse 1980, 2) & Wiesbaden:
Franz Steiner Verlag.
Christmann H H & Hausmann F-R (eds.) (1989). In
Verbindung mit Manfred Briegel (1989). Deutsche
und osterreichische Romanisten als Verfolgte des Nationalsozialismus. Mit einer bio-bibliographischen Dokumentation der verfolgten Romanisten (Romanica et
Comparatistica, vol. 10). Tubingen: Stauffenburg.
Meyer-Lubke W (18901902). Grammatik der romanischen Sprachen (4 vols). Leipzig: Fues/Reisland.
Rickford, John R. (b. 1949) 627

Pulgram E (1996). Elise Richter (18651943). In
Stammerjohann H et al. (eds.) Lexicon Grammaticorum.
Whos Who in the History of World Linguistics. Tu bingen: Max Niemeyer Verlag. 791792.
Renzi L (1987). Elise Richter (18651943). Studi Goriziani 65, 99111.
Richter E (1903). Zur Entwicklung der romanischen
Wortstellung aus der lateinischen. Halle: Niemeyer.
Richter E (1904). Ab im Romanischen. Halle: Niemeyer.
Richter E (1909). Die Rolle der Semantik in der historischen Grammatik. In Germanisch-Romanische Monatszeitschrift (GRM). Bd. 1 (1909) 5 (1913) bearbeitet
von Nadine Krolla, GRM 2 (1910), S. 231243.
Richter E (1911). Der innere Zusammenhang in der
Entwicklung der romanischen Sprachen. In Prinzipien-
fragen der romanischen Sprachwissenschaft. Wilhelm

Meyer-Lu bke zur Feier der Vollendung seines 50. Lebensjahres gewidmet, Teil II: 57143. Beihefte zur Zeitschrift
fu r romanische Philologie. Hrsg. v. Gustav Groeber.
Halle/Saale: Niemeyer
Richter E (1977). Kleinere Schriften zur Allgemeinen und
Romanischen Sprachwissenschaft. Ausgewa hlt, eingeleitet und kommentiert von Yakov Malkiel. Mit einer Bibliographie von B. M. Woodbridge, J R. Gesamtredaktion:
Wolfgang Meid. Innsbrucker Beitrage zur Sprachwissenschaft. Innsbruck: H. Kowatsch.
Woodbridge B M Jr (1972). A bibliography of the writings
of Elise Richter. Romance Philology 26, 342360.
Rickford, John R. (b. 1949)

B McElhinny, University of Toronto, Ontario, Canada
Dr. John Rickford is Martin Luther King Jr. Centennial

Professor of Linguistics and director of African and
AfroAmerican Studies at Stanford University. He
was born in Georgetown, Guyana, and has degrees
from the University of California Santa Cruz (B.A.,
1971) and the University of Pennsylvania (M.A.,
1973; Ph.D. 1979). His work focuses on the history
and current use of Caribbean Creoles and of African
American Vernacular English (AAVE), with particular
emphasis on the development of variationist sociolinguistic methods, debates about the creolist and anglicist
roots of AAVE, and the educational implications of
linguistic analysis (the latter often in conjunction with
his wife, Dr. Angela Rickford). Other important contributions have included the call for new approaches to
class analysis in sociolinguistics, studies of ethnicity as
a sociolinguistic barrier, sociolinguistic analyses of
style, and articles about the ethical obligations of social
scientists. He has published 10 books and over 70
articles, some of the most influential of which are listed
in the following bibliography.
In 1996, after the Oakland school boards widely
and heatedly discussed decision to recognize the vernacular of AfricanAmerican students in an Ebonics
instruction program, Dr. Rickford engaged in countless public discussions with academics, teachers,
churches, professional associations, and journalists
on the ways that widespread knowledge of the structure and history of AAVE can improve the success of
AfricanAmerican students in the language arts.
This work was recognized by the Linguistic Society
of America with its Linguistics, Language, and the

Public award, and by the American Anthropology
Association with its Anthropology in Media award.
Spoken soul: the story of black English (2000), a
book written with his son, journalist Russell John
Rickford, demonstrates his flair for offering an accessible introduction to AAVE, and also includes
an insightful overview of the debates around the
Oakland school board decision. It is the 2002
winner of an American Book Award. A talented and
award-winning teacher, John Rickford is known for
exciting the interest and developing the skills of undergraduate and graduate students by creating numerous opportunities for them to engage in original,
collaborative research, and for devoting generous
amounts of time to mentoring junior scholars.
See also: English, African American Vernacular; Pidgins
and Creoles: Overview; Sociolinguistics and Political
Economy.
Bibliography
Rickford J R (1974). The insights of the Mesolect. In
DeCamp D & Hancock I (eds.) Pidgins and creoles.
Washington DC: Georgetown University Press.
Rickford J R (1977). The question of prior creolization in
black English. In Valdman A (ed.) Pidgin and creole
linguistics. Bloomington, Indiana: Indiana University
Press.
Rickford J R (1985). Ethnicity as a sociolinguistic boundary. American Speech 60(3), 90125.
Rickford J R (1986). The need for new approaches to
social class analysis in sociolinguistics. Language and
Communication 6(3), 215221.
Rickford, John R. (b. 1949) 627

Pulgram E (1996). Elise Richter (18651943). In
Stammerjohann H et al. (eds.) Lexicon Grammaticorum.
Whos Who in the History of World Linguistics. Tubingen: Max Niemeyer Verlag. 791792.
Renzi L (1987). Elise Richter (18651943). Studi Goriziani 65, 99111.
Richter E (1903). Zur Entwicklung der romanischen
Wortstellung aus der lateinischen. Halle: Niemeyer.
Richter E (1904). Ab im Romanischen. Halle: Niemeyer.
Richter E (1909). Die Rolle der Semantik in der historischen Grammatik. In Germanisch-Romanische Monatszeitschrift (GRM). Bd. 1 (1909) 5 (1913) bearbeitet
von Nadine Krolla, GRM 2 (1910), S. 231243.
Richter E (1911). Der innere Zusammenhang in der
Entwicklung der romanischen Sprachen. In Prinzipien-
fragen der romanischen Sprachwissenschaft. Wilhelm

Meyer-Lubke zur Feier der Vollendung seines 50. Lebensjahres gewidmet, Teil II: 57143. Beihefte zur Zeitschrift
fur romanische Philologie. Hrsg. v. Gustav Groeber.
Halle/Saale: Niemeyer
Richter E (1977). Kleinere Schriften zur Allgemeinen und
Romanischen Sprachwissenschaft. Ausgewahlt, eingeleitet und kommentiert von Yakov Malkiel. Mit einer Bibliographie von B. M. Woodbridge, J R. Gesamtredaktion:
Wolfgang Meid. Innsbrucker Beitrage zur Sprachwissenschaft. Innsbruck: H. Kowatsch.
Woodbridge B M Jr (1972). A bibliography of the writings
of Elise Richter. Romance Philology 26, 342360.
Rickford, John R. (b. 1949)

B McElhinny, University of Toronto, Ontario, Canada
Dr. John Rickford is Martin Luther King Jr. Centennial

Professor of Linguistics and director of African and
AfroAmerican Studies at Stanford University. He
was born in Georgetown, Guyana, and has degrees
from the University of California Santa Cruz (B.A.,
1971) and the University of Pennsylvania (M.A.,
1973; Ph.D. 1979). His work focuses on the history
and current use of Caribbean Creoles and of African
American Vernacular English (AAVE), with particular
emphasis on the development of variationist sociolinguistic methods, debates about the creolist and anglicist
roots of AAVE, and the educational implications of
linguistic analysis (the latter often in conjunction with
his wife, Dr. Angela Rickford). Other important contributions have included the call for new approaches to
class analysis in sociolinguistics, studies of ethnicity as
a sociolinguistic barrier, sociolinguistic analyses of
style, and articles about the ethical obligations of social
scientists. He has published 10 books and over 70
articles, some of the most influential of which are listed
in the following bibliography.
In 1996, after the Oakland school boards widely
and heatedly discussed decision to recognize the vernacular of AfricanAmerican students in an Ebonics
instruction program, Dr. Rickford engaged in countless public discussions with academics, teachers,
churches, professional associations, and journalists
on the ways that widespread knowledge of the structure and history of AAVE can improve the success of
AfricanAmerican students in the language arts.
This work was recognized by the Linguistic Society
of America with its Linguistics, Language, and the

Public award, and by the American Anthropology
Association with its Anthropology in Media award.
Spoken soul: the story of black English (2000), a
book written with his son, journalist Russell John
Rickford, demonstrates his flair for offering an accessible introduction to AAVE, and also includes
an insightful overview of the debates around the
Oakland school board decision. It is the 2002
winner of an American Book Award. A talented and
award-winning teacher, John Rickford is known for
exciting the interest and developing the skills of undergraduate and graduate students by creating numerous opportunities for them to engage in original,
collaborative research, and for devoting generous
amounts of time to mentoring junior scholars.
See also: English, African American Vernacular; Pidgins
and Creoles: Overview; Sociolinguistics and Political
Economy.
Bibliography
Rickford J R (1974). The insights of the Mesolect. In
DeCamp D & Hancock I (eds.) Pidgins and creoles.
Washington DC: Georgetown University Press.
Rickford J R (1977). The question of prior creolization in
black English. In Valdman A (ed.) Pidgin and creole
linguistics. Bloomington, Indiana: Indiana University
Press.
Rickford J R (1985). Ethnicity as a sociolinguistic boundary. American Speech 60(3), 90125.
Rickford J R (1986). The need for new approaches to
social class analysis in sociolinguistics. Language and
Communication 6(3), 215221.
628 Rickford, John R. (b. 1949)

Rickford J R (1987). Are black and white vernaculars
diverging? American Speech 62(1), 5562.
Rickford J R (1987). Dimensions of a Creole continuum.
Stanford: Stanford University Press.
Rickford J R (1997). Unequal partnership: sociolinguistics and the African American speech community.
Language in Society 26, 161197.
Rickford J R (1998). The Creole origins of African
American vernacular English. African American
English. London: Routledge.
Rickford J R (1999). African American vernacular
English: features and use, evolution, and educational
implications. Oxford: Blackwell.
Rickford J R & Eckert P (2002). Style and sociolinguistic

variation. Cambridge: Cambridge University Press.
Rickford J R & Finegan E (2004). Language in the USA:
themes for the new millennium. Cambridge: Cambridge
University Press.
Rickford J R, Mufwene S, Bailey G & Baugh J (eds.) (1998).
African American English. London: Routledge.
Rickford J R & Rickford A (1976). Cut-eye and suck-teeth:
African words and gestures in New World guise. Journal
of American Folklore 89, 194209.
Rickford J R & Rickford R J (2000). Spoken soul: the story
of black English. New York: John Wiley.
Right Node Raising
Right node raising (RNR), a term due to Postal

(1974), describes an analysis that assumes the target,
which must be phonetically and grammatically identical in both conjuncts, to raise across the board
(ATB) to a right-adjoined position above the coordination. Raising is indicated by coindexed traces (t) in
the base position (Figure 1). This analysis of RNR,
which is found in Postal (1974, 1998), Williams
(1990), Larson (1990), and Sabbagh (2003), among

others, is referred to here as the raising analysis.
An alternative analysis of RNR dispenses with
literal raising of the target and assumes that the target
is syntactically present in its base position in both
conjuncts, but is phonetically empty in the first one
(Figure 2, the phonetic analysis; see Wexler and Culicover (1980), Wilder (1997), and Hartmann (2000)).
A third possibility to analyze RNR is to allow the
target to be multiply dominated by two nodes at the
same time. In this view (Figure 3), the target occupies
several syntactic positions but is pronounced only
in the rightmost position of the clause; this is the
analysis of McCawley (1982, 1988), Blevins (1990),
and Steedman (1985).
There has been a long-standing debate in the linguistic literature as to which of the three analyses is
the most appropriate one. In the following discussions, some of the arguments that have been brought
up to strengthen the phonetic analysis, in preference
to the raising analysis, are given. This presentation
does not intend to reproduce all aspects of the still
ongoing debate; rather, it merely aims to give an
impression of the kinds of constructions and arguments to be considered when investigating RNR.
Figure 1 The raising analysis.
Figure 2 The phonetic analysis.
K Hartmann, Humboldt-Universitatzu Berlin, Berlin,

Germany
The term right node raising refers to a coordination

construction in which the right end of the first conjunct (i.e., the target) is not expressed, creating an
information gap. The gap is resolved at the end of the
last conjunct, where the missing part of the first conjunct is expressed overtly, as shown in Example (1).
Here and in all subsequent examples, the target
appears in square brackets; the information gap in
the first conjunct is indicated by an underline ( ):
(1) Sarah baked , and Louise devoured [five
pancakes].
628 Rickford, John R. (b. 1949)

Rickford J R (1987). Are black and white vernaculars
diverging? American Speech 62(1), 5562.
Rickford J R (1987). Dimensions of a Creole continuum.
Stanford: Stanford University Press.
Rickford J R (1997). Unequal partnership: sociolinguistics and the African American speech community.
Language in Society 26, 161197.
Rickford J R (1998). The Creole origins of African
American vernacular English. African American
English. London: Routledge.
Rickford J R (1999). African American vernacular
English: features and use, evolution, and educational
implications. Oxford: Blackwell.
Rickford J R & Eckert P (2002). Style and sociolinguistic

variation. Cambridge: Cambridge University Press.
Rickford J R & Finegan E (2004). Language in the USA:
themes for the new millennium. Cambridge: Cambridge
University Press.
Rickford J R, Mufwene S, Bailey G & Baugh J (eds.) (1998).
African American English. London: Routledge.
Rickford J R & Rickford A (1976). Cut-eye and suck-teeth:
African words and gestures in New World guise. Journal
of American Folklore 89, 194209.
Rickford J R & Rickford R J (2000). Spoken soul: the story
of black English. New York: John Wiley.
Right Node Raising
Right node raising (RNR), a term due to Postal

(1974), describes an analysis that assumes the target,
which must be phonetically and grammatically identical in both conjuncts, to raise across the board
(ATB) to a right-adjoined position above the coordination. Raising is indicated by coindexed traces (t) in
the base position (Figure 1). This analysis of RNR,
which is found in Postal (1974, 1998), Williams
(1990), Larson (1990), and Sabbagh (2003), among

others, is referred to here as the raising analysis.
An alternative analysis of RNR dispenses with
literal raising of the target and assumes that the target
is syntactically present in its base position in both
conjuncts, but is phonetically empty in the first one
(Figure 2, the phonetic analysis; see Wexler and Culicover (1980), Wilder (1997), and Hartmann (2000)).
A third possibility to analyze RNR is to allow the
target to be multiply dominated by two nodes at the
same time. In this view (Figure 3), the target occupies
several syntactic positions but is pronounced only
in the rightmost position of the clause; this is the
analysis of McCawley (1982, 1988), Blevins (1990),
and Steedman (1985).
There has been a long-standing debate in the linguistic literature as to which of the three analyses is
the most appropriate one. In the following discussions, some of the arguments that have been brought
up to strengthen the phonetic analysis, in preference
to the raising analysis, are given. This presentation
does not intend to reproduce all aspects of the still
ongoing debate; rather, it merely aims to give an
impression of the kinds of constructions and arguments to be considered when investigating RNR.
Figure 1 The raising analysis.
Figure 2 The phonetic analysis.
K Hartmann, Humboldt-Universitatzu Berlin, Berlin,

Germany
The term right node raising refers to a coordination

construction in which the right end of the first conjunct (i.e., the target) is not expressed, creating an
information gap. The gap is resolved at the end of the
last conjunct, where the missing part of the first conjunct is expressed overtly, as shown in Example (1).
Here and in all subsequent examples, the target
appears in square brackets; the information gap in
the first conjunct is indicated by an underline ( ):
(1) Sarah baked , and Louise devoured [five
pancakes].
Right Node Raising 629
Considered in turn are the constituency of the target,

the absence of locality restrictions in RNR, the interpretation of the alleged traces, and the scope effects
between elements of the target and the remnant
conjuncts.
The first argument in favor of the phonetic analysis
and against the raising analysis concerns constituency. Movement (including rightward movement) is
generally restricted to one syntactic constituent. The
target in RNR, however, can either contain several
constituents, i.e., a noun phrase (NP) and a prepositional phrase (PP), as in Example (2a) (Abbott, 1976),
or it can be a nonconstituent, e.g., as the head of a
noun phrase (Example (2b)). This deviation from
regular movement has led several researchers to the
assumption that RNR cannot be raising (e.g., see
Gleitman, 1965; Abbott, 1976; Williams, 1978; van
Oirsouw, 1983, 1987).
(2a) Smith loaned , and his widow later donated
[a valuable collection of manuscripts to the
library].
(2b) Louise likes yellow , but Sarah prefers pink
[shirts].
The following German examples also show clearly

that RNR does not respect constituency. In German,
prepositions may not stay behind if the complement
of the preposition is displaced. This holds for leftward
movement of the complement (Example (3a)), as well
as for rightward movement (Example (3b)). Example
(3b) illustrates heavy-NP-shift of a noun phrase originating in the complement position of the preposition
uber about. The ungrammaticality of both examples
indicates that German disallows preposition stranding. As shown in Example (3c), heavy-NP-shift is
possible if no preposition is stranded in the midfield
(t, trace of movement; DAT, dative):
(3a) *[NPDem Tisch]1 sitzt Louise [PPauf t1].
the.DAT
table
sits
Louise on
(3b) *Auf dem Empfang wurden wir
at
the reception
were
we
[PPu ber t1] informiert [NPdas
about
informed
the
Figure 3 The multidominance analysis.
sehr reichhaltige Angebot

an erlesenen
very varied
collection of selected
Speisen und Getra nken]1.
dishes
and drinks
(3c) Auf dem Empfang wurde uns t1 serviert
at the reception was
us.DAT served
[NP ein sehr reichhaltiges Angebot an erlesenen
a very varied
collection of selected
Speisen und Getra nken]1.
dishes and drinks
At the reception, we were served a very varied
collection of selected dishes and drinks.
As Example (4) shows, a preposition may strand in

German RNR (Hartmann, 2000: 58), and the same
holds for Irish (see McCloskey, 1986). This fact
strongly suggests that RNR does not involve rightward ATB movement of the target:
(4) Louise sitzt auf , und Sarah sitzt unter
Louise sits on
and Sarah sits under
[dem Tisch].
the
table
Louise is sitting on, and Sarah is sitting under the
table.
The second argument corroborating the phonetic

analysis shows that RNR does not obey well-known
locality restrictions. It has been shown that targets
may be contained within syntactic islands. Wexler
and Culicover (1980), Goodall (1987), and Phillips
(1996) argued that RNR freely violates the whconstraint: in Example (5a) (Wexler and Culicover,
1980: 299), the target is contained in a wh-clause,
which usually represents an island for movement.
Neijt (1979: 44) claimed that the target may be part
of a complex NP (Example (5b)), known to disallow
extraction of constituents contained therein. Thus,
RNR does not seem to be restricted by typical locality
constraints:
(5a) Mary knows a man [Swho buys ], and Bill
knows a man who sells [pictures of Fred].
(5b) Alfonse discussed [NPthe question of which
rice ], and Mary discussed the question of
which beans [we would eat].
The references compare RNR only with leftward

movement, but since movement to the right considerably differs from movement to the left, the comparison must include RNR and clear instances of
rightward movement, in order to make the argument
valid. It can be argued that extraposition ( movement to the right) may not violate the coordinate
structure constraint. In Example (6a), the PP to the
swimming pool is extraposed out of the first conjunct
of a coordinated verb phrase (VP), resulting in an
ungrammatical sentence. In RNR (Example (6b)),
on the other hand, the coordinate structure constraint
630 Right Node Raising
does not seem to be violated: although the PP to

school originates in a coordinated VP, it can function
as an RNR target. This would be unexpected if RNR
were an instance of rightward ATB movement:
(6a) *Louise [VP[VPgoes t1] and [VPreads a book]] [PPto
the swimming pool]1.
(6b) Louise [VP[VPhad coffee] and [VPwalked ]], but
Sarah overslept and ran [to school].
The third argument is semantic in nature. It concerns the interpretation of traces. If RNR did involve
movement, the traces of the alleged raising operation
would be expected to be referentially identical. In
other words, the target should refer to one and the
same referent in both conjuncts. As Example (7b)
shows, this expectation is not borne out (the argument
is made for German in Klein (1993)):
(7a) What1 did Louise buy t1 and Sarah read t1?
(7b) Louise bought , and Sarah read [a novel].
In Example (7a), the wh-object is moved across the

board to the sentence-initial position. The unique
interpretation is that Louise and Sarah bought and
read the same item. This reading is not predominant
in the RNR sentence (Example (7b)), in which the
novel bought by Louise and the novel read by Sarah
may differ. Referential identity of the target is clearly
excluded in Example (8); here, six goals were scored
altogether, not three (unless they played against each
other, of course!):
(8) Real Madrid scored , and Bayern Mu nchen got
[three goals].
Since ATB movement obviously triggers referential

identity of the moved constituents in both conjuncts,
RNR in Examples (7b) and (8) cannot be derived by
ATB movement, as assumed by the raising analysis.
The phonetic analysis, on the other hand, captures the
facts correctly.
The final argument presented here concerns scope
effects. The raising analysis wrongly predicts that the
target should be outside the scope of any element
contained in the remnant conjuncts. However, the
target appears not to have moved when variable binding, negative polarity licensing, and binding theory
are considered (see Levine, 1985; Phillips, 1996;
Hartmann, 2000). Phillips (1996: 53) argued that a
quantifier may bind a pronoun contained in a target.
This is possible only if the target is not attached
higher in the tree than the quantifier is, given that
the quantifier must take scope over the pronoun in
order to bind it. As indicated by the indices, a bound
reading of the pronoun is possible:
(9) Everyone1 liked , and at least one person1 loved
[the paper he1 had been asked to review].
It is known that negative polarity items must appear

in the scope of some negative element. According to
the raising analysis, a negative polarity item in a
target should be excluded, since the raised target is
outside of the scope of the coordinated structure. This
expectation is not borne out, as illustrated by Example (10) (Phillips, 1996: 53):
(10) Nobody enjoyed , and few people even liked
[any of the talks on RNR].
Finally, Levine (1985: 496) pointed out that the ungrammaticality of Example (11) cannot be explained
if the target raises to some right-peripheral position,
thereby bleeding the illicit binding of the referential
expression Mary (contained in the target) by the
pronoun she. The phonetic analysis offers a natural
account for the ungrammaticality of Example (11),
assuming that the target does not move:
(11) *I know that she1 said and I happen to agree
[that Mary1 needs a new car].
To summarize, the aforementioned arguments all

corroborate the phonetic analysis, which assumes
that the target does not move, as first noted by Wexler
and Culicover (1980). As has been indicated herein,
these arguments represent only some aspects of an
interesting and still open debate.
See also: Binding Theory; Coordination; Island Con-
straints.
Bibliography
Abbott B (1976). Right node raising as a test for constituenthood. Linguistic Inquiry 7, 639642.
Blevins J (1990). Syntactic complexity: evidence for discontinuity and multidomination. Ph.D. thesis, University of
Massachusetts at Amherst.
Fe ry C & Hartmann K (2004). The focus and prosodic
structure of German right node raising and gapping.
The Linguistic Review 22, 67114.
Gleitman L (1965). Coordinating conjunctions in English.
Language 41, 260293.
Goodall G (1987). Parallel structures in syntax. Coordination, causatives and restructuring. Cambridge: Cambridge
University Press.
Hartmann K (2000). Right node raising and gapping.
Interface conditions on prosodic deletion. Amsterdam/
Philadelphia: John Benjamins.
Johannessen J B (1998). Coordination. Oxford: Oxford
University Press.
Klein W (1993). Ellipse. In Jacobs J et al. (eds.) Syntax
Ein internationales Handbuch zeitgeno ssischer Forschung. 1. Halbband. Berlin/New York: Walter de
Gruyter. 763799.
Rigid Designation 631

Larson R (1990). Double objects revisited: reply to
Jackendoff. Linguistic Inquiry 21, 589632.
Levine R (1984). A note on right node raising, tough
constructions and reanalysis rules. Linguistic Analysis
13, 159172.
Levine R (1985). Right node (non-) raising. Linguistic
Inquiry 16, 492497.
McCawley J (1982). Parentheticals and discontinuous
constituents. Linguistic Inquiry 13, 91106.
McCawley J (1988). The syntactic phenomena of English.
Chicago: Chicago University Press.
McCloskey J (1986). Right node raising and preposition
stranding. Linguistic Inquiry 17, 183186.
Munn A (1993). Topics in the syntax and semantics of coordinate structures. Ph.D. thesis, University of Maryland.
Neijt A (1979). Gapping. Dordrecht: Foris.
Phillips C (1996). Order and structure. Ph.D. thesis,
Massachusetts Institute of Technology.
Postal P (1974). On raising. Cambridge, MA: MIT Press.
Postal P (1998). Three investigations of extraction.
Cambridge, MA: MIT Press.
Sabbagh J (2003). Right node raising is sometimes movement before and sometimes movement after spellout.
Paper presented at West Coast Conference on Formal

Linguistics 22.
Steedman M (1985). Dependency and coordination in the
grammar of Dutch and English. Language 61, 523568.
Swingle K (1994). The role of prosody in right node raising. In Pullum G K & Potsdam E (eds.) Syntax at Santa
Cruz, vol. 2. Santa Cruz, CA: UCSC Linguistics Research
Center.
van Oirsouw R (1983). Coordinated sentences. Lingua 60,
135145.
van Oirsouw R (1987). The syntax of coordination.
London: Croom Helm.
Wexler K & Culicover P (1980). Formal principles of language acquisition. Cambridge, MA: MIT Press.
Wilder C (1997). Some properties of ellipsis in coordination. In Alexiadou A & Hall T A (eds.) Studies on universal grammar and typological variation. Amsterdam/
Williams E (1978). Across-the-board rule application.
Williams E (1990). The ATB theory of parasitic gaps. The
Linguistic Review 6, 265279.
Rigid Designation
S Predelli, University of Nottingham, Nottingham, UK
Introduction
Rigidity is a semantic notion. Semantics studies the
relationships between expressions in a language and
(typically) extralinguistic items. Examples of semantically interesting features are the relationship
between the name George Bush and Bush, the individual to which it refers; the relationship between the
sentence George Bush is a liberal and falsehood, its
truth value; and, at least in some views, the relationship between the definite description the President of
the United States and Bush, the individual it denotes.
The task of a semantic theory is not that of immediately assigning the aforementioned features to the
appropriate expressions. For instance, that the sentence George Bush is a liberal is to be assigned
falsehood depends not only on the semantic traits of
the expressions under analysis, but also on obviously
extrasemantic facts having to do with Bushs political
convictions. By the same token, that the denotatum of
the President of the United States is George Bush,
rather than John Kerry or myself, also depends on
extrasemantic facts having to do with the results
of the latest presidential election. For this reason,
the aim of a semantic theory is that of presenting
relativized assignments of semantic features. For instance, what our theory may conclude is that George
Bush is a liberal is true with respect to all and only
those parameters that provide a positive reply to the
query about Bushs leftist tendencies, and that the
President of the United States denotes Bush with
respect to all and only those parameters with respect
to which Bush won the presidential elections. According to a customary approach, parameters of this type
may be understood as possible worlds. In a more
complex framework, the relata appropriate for semantic evaluation may include, together with possible worlds, a temporal parameter: for instance, the
President of the United States may denote Bush with
respect to a given possible world and a certain time,
but Bill Clinton with respect to that world and another time span.
Let me refer to the semantic features for the aforementioned expressions as their semantic value: so, the
semantic value of a sentence (with respect to a possible world and/or a time) is its truth value (with respect to that world and that time); the semantic value
of a definite description (with respect to a possible
world and/or a time) is its denotatum (with respect to
that world and that time); and so on. In general, and
leaving momentarily aside a few interesting complications, an expression is rigid with respect to a parameter k if it has the same semantic value across all
Rigid Designation 631

Larson R (1990). Double objects revisited: reply to
Jackendoff. Linguistic Inquiry 21, 589632.
Levine R (1984). A note on right node raising, tough
constructions and reanalysis rules. Linguistic Analysis
13, 159172.
Levine R (1985). Right node (non-) raising. Linguistic
Inquiry 16, 492497.
McCawley J (1982). Parentheticals and discontinuous
constituents. Linguistic Inquiry 13, 91106.
McCawley J (1988). The syntactic phenomena of English.
Chicago: Chicago University Press.
McCloskey J (1986). Right node raising and preposition
stranding. Linguistic Inquiry 17, 183186.
Munn A (1993). Topics in the syntax and semantics of coordinate structures. Ph.D. thesis, University of Maryland.
Neijt A (1979). Gapping. Dordrecht: Foris.
Phillips C (1996). Order and structure. Ph.D. thesis,
Massachusetts Institute of Technology.
Postal P (1974). On raising. Cambridge, MA: MIT Press.
Postal P (1998). Three investigations of extraction.
Cambridge, MA: MIT Press.
Sabbagh J (2003). Right node raising is sometimes movement before and sometimes movement after spellout.
Paper presented at West Coast Conference on Formal

Linguistics 22.
Steedman M (1985). Dependency and coordination in the
grammar of Dutch and English. Language 61, 523568.
Swingle K (1994). The role of prosody in right node raising. In Pullum G K & Potsdam E (eds.) Syntax at Santa
Cruz, vol. 2. Santa Cruz, CA: UCSC Linguistics Research
Center.
van Oirsouw R (1983). Coordinated sentences. Lingua 60,
135145.
van Oirsouw R (1987). The syntax of coordination.
London: Croom Helm.
Wexler K & Culicover P (1980). Formal principles of language acquisition. Cambridge, MA: MIT Press.
Wilder C (1997). Some properties of ellipsis in coordination. In Alexiadou A & Hall T A (eds.) Studies on universal grammar and typological variation. Amsterdam/
Williams E (1978). Across-the-board rule application.
Williams E (1990). The ATB theory of parasitic gaps. The
Linguistic Review 6, 265279.
Rigid Designation
S Predelli, University of Nottingham, Nottingham, UK
Introduction
Rigidity is a semantic notion. Semantics studies the
relationships between expressions in a language and
(typically) extralinguistic items. Examples of semantically interesting features are the relationship
between the name George Bush and Bush, the individual to which it refers; the relationship between the
sentence George Bush is a liberal and falsehood, its
truth value; and, at least in some views, the relationship between the definite description the President of
the United States and Bush, the individual it denotes.
The task of a semantic theory is not that of immediately assigning the aforementioned features to the
appropriate expressions. For instance, that the sentence George Bush is a liberal is to be assigned
falsehood depends not only on the semantic traits of
the expressions under analysis, but also on obviously
extrasemantic facts having to do with Bushs political
convictions. By the same token, that the denotatum of
the President of the United States is George Bush,
rather than John Kerry or myself, also depends on
extrasemantic facts having to do with the results
of the latest presidential election. For this reason,
the aim of a semantic theory is that of presenting
relativized assignments of semantic features. For instance, what our theory may conclude is that George
Bush is a liberal is true with respect to all and only
those parameters that provide a positive reply to the
query about Bushs leftist tendencies, and that the
President of the United States denotes Bush with
respect to all and only those parameters with respect
to which Bush won the presidential elections. According to a customary approach, parameters of this type
may be understood as possible worlds. In a more
complex framework, the relata appropriate for semantic evaluation may include, together with possible worlds, a temporal parameter: for instance, the
President of the United States may denote Bush with
respect to a given possible world and a certain time,
but Bill Clinton with respect to that world and another time span.
Let me refer to the semantic features for the aforementioned expressions as their semantic value: so, the
semantic value of a sentence (with respect to a possible world and/or a time) is its truth value (with respect to that world and that time); the semantic value
of a definite description (with respect to a possible
world and/or a time) is its denotatum (with respect to
that world and that time); and so on. In general, and
leaving momentarily aside a few interesting complications, an expression is rigid with respect to a parameter k if it has the same semantic value across all
632 Rigid Designation
ks. For instance, an eternally true sentence is temporally rigid: its truth value is constant across all times.
Since in what follows I leave temporal considerations
out of the picture, I employ the expression rigid tout
court as alluding to constancy of semantic value
across possible worlds.
Names and Rigidity

Two kinds of singular designators, namely names and
definite descriptions, have been singled out as paradigmatic of the rigid/nonrigid divide. As I hinted, typical definite descriptions such as the President of the
United States are nonrigidly associated with distinct
denotata with respect to different possible worlds.
But proper names such as George Bush apparently
rigidly persist in referring to Bush with respect to all
possible worlds (or, at least, all possible worlds in
which Bush exists; more on this caveat later). For
instance, if you are interested in evaluating George
Bush is wise with respect to a counterfactual scenario
as dramatically different from the actual world as you
may like, what you need to consider is whether, given
the collection of individuals that are wise in that
scenario, Bush is one of them (see Kripke, 1980).
One possible source of confusion must be cleared
from the outset: what is at issue is not the question of
who happens to be called George Bush in the scenarios under consideration: that in some merely possible world certain expressions may be endowed with
a different meaning, and certain names may be used
to refer to alternative individuals is of no relevance
for our purpose. Similarly irrelevant is the possibility
that, in our actual employment of George Bush,
different individuals may be referred to by that
name type.
Types of Rigidity
I wrote that typical definite descriptions behave nonrigidly. This does not entail that there may not be
rigid definite descriptions. For instance, the positive
square root of four denotes the number two with
respect to any possible world and is thus a rigid
designator. It seems clear that this expressions rigidity derives not directly from its semantic behavior, but
from the modal perseverance of the mathematical
world. The description the positive square root of
four interacts with the parameter with respect to
which it is evaluated as any other description would:
a certain condition is being put forth, in this case that
of being the positive square root of four, and an
individual is selected on the basis of its ability to
satisfy that condition. The description is, in a sense,
semantically available for a nonrigid profile. Its
rigidity is a result not of its semantic behavior, but of

the response that alternative possible worlds provide:
one and the same item, the number two, is inevitably
being offered as the descriptions denotatum.
Saul Kripke refers to descriptions such as the positive square root of four as rigid only de facto. Expressions such as proper names, namely expressions
whose semantic profile presumably guarantees their
rigidity, are classified as rigid de jure (Kripke, 1980).
Note that in this understanding of the de facto/de jure
distinction, not all rigid descriptions are rigid merely
de facto. For instance, the description the actual
President of the United States rigidly refers to a certain individual by virtue of its semantic behavior,
in particular by virtue of the meaning of the indexical
expression actual (see Kaplan, 1977; Lewis, 1970).
Thus far, I rested satisfied with the notion that a
rigid expression has a constant semantic value across
possible worlds, and, in particular, that a rigid designator designates the same individual with respect
to any possible world in which that object exists
(Kripke, 1980; see also Kaplan, 1977). When it
comes to the decision about an expressions semantic
value with respect to possible worlds in which that
object fails to exist, two options may be entertained.
In the terminology from Salmon (1981), a designator
is persistently rigid iff it designates its designatum x
with respect to any world in which x exists and designates nothing with respect to any world in which x
does not exist. On the other hand, a designator is
obstinately rigid iff it designates x with respect to
every possible world, regardless of whether x does
or does not exist in it.
Rigidity has played an important role in one of the
central semantic debates during the seventies and
eighties, having to do with descriptivist views of customary proper names, such as, presumably, Freges
and Russells. If proper names are rigid but typical
descriptions are not, so the story goes, proper names
may not behave in the manner descriptivists envisage.
(For a critical discussion of a descriptivist reply, to the
effect that proper names are associated with rigid
descriptions, see Salmon, 1981; Soames, 1998.)
Other important themes linked to the notion of rigidity have to do with the hypothesis that natural kind
terms, such as tiger or gold, also behave rigidly (see
Kripke, 1980; for a detailed discussion of the difficulties in extending the classic definition of rigidity to
kind terms, see Soames, 2002). Finally, the idea of
rigidity and the distinct but related notion of direct
reference occupy center stage in the semantic analysis of indexical expressions. (On the relationships
between rigidity and direct reference, see Marti,
2003.) For instance, an indexical singular term such
as I rigidly refers (with respect to a given context c) to
Ritual and Religious Language 633
the agent of c, that is, at least in typical instances,

to the person who is speaking. The notion that indexicals such as I behave rigidly, even though they are
associated with a descriptive meaning, plays a crucial
role in the classic arguments for the distinction between meaning and content or, in David Kaplans
classic vocabulary, between character and content
(see Kaplan, 1977; Lewis, 1980).
See also: Descriptions, Definite and Indefinite: Philosophical Aspects; Indexicality: Theory; Modal Logic; Possible
Worlds: Philosophical Theories.
Lewis D (1970). Anselm and actuality. Nous 4, 175188.

Kripke S (1980). Naming and necessity (2nd edn.).
Cambridge, MA: Harvard University Press.
Lewis D (1980). Index, context, and content. In Kanger S
hman S (eds.) Philosophy and grammar. Dordrecht:
&O
Reidel.
Marti G (2003). A question of rigidity in new theories of
reference. Nous 37, 161179.
Salmon N (1981). Reference and essence. Princeton, NJ:
Princeton University Press.
Soames S (1998). The modal argument: wide scope and
rigidified descriptions. Nous 32, 122.
Soames S (2002). Beyond rigidity. The unfinished semantic
agenda of naming and necessity. Oxford: Oxford University Press.
Bibliography
Kaplan D (1977). Demonstratives. In Almog J, Perry J &
Wettstein H (eds.) Themes from Kaplan. New York and
Oxford: Oxford University Press.
Ritual and Religious Language

R A Yelle, University of Toronto, Toronto,
Ontario, Canada
Definition and Varieties

The concept of ritual language presupposes a concept
of ritual, a category beset by numerous definitional
problems. At times it appears that the category of
ritual has become a holding-area for all those forms
of behavior to which we cannot attribute any utilitarian value. This has led both to confusion over the
definition of ritual, and to a common pejorative evaluation of ritual as meaningless, as illustrated by the
modern history of the theorization of this category.
There is a line of influence that extends from Reformation polemics against traditional Catholic liturgical
practices to 19th- and early 20th-century anthropologies characterizing magic as a survival from primitive
culture (Tylor, 1903), as well as a mistaken theory and
praxis outmoded by science (Frazer, 1951). Later sociological interpretations recovered a positive value
for rituals by emphasizing the role that these play in
constructing, reinforcing, and contesting the dynamics of social groups (Durkheim, 1965; Gluckman,
1954; Turner, 1977; Bourdieu, 1977; Lincoln, 1994).
Due to their emphasis on social dynamics, and despite
their frequent application of structuralist analysis to
ritual processes, such interpretations have generally
paid little attention to the specific forms of ritual
language. However, other scholars have emphasized

the performative or pragmatic dimensions of ritual
(Tambiah, 1985), and occasionally conducted sophisticated semiotic analyses of ritual language (Sebeok,
1964; Silverstein and Urban, 1996). Taking all of these
approaches into account produces a more coherent
theory of ritual language.
There exist numerous terms for different types of
ritual language. Each culture of course has its own
terms. Common English terms are spell, charm,
incantation, chant, prayer, and glossolalia.
More recent, theoretical terms include performative
utterance (Austin, 1975), speech act (Searle, 1969),
and indexical icon (Silverstein, 1993, 1998; Silverstein and Urban, 1996). Although by no means exhaustive, this list is representative, and it indicates
some of the important variations among types of
ritual language. Spells and chants are often formulaic
and invariant, whereas glossolalia and certain forms
of prayer are, at least ostensibly, spontaneous and
improvised. Some forms of ritual language aim to
achieve obviously practical objectives, whereas
others, including some forms of prayer, do not. Historically, these very distinctions have informed some
attempts to distinguish good rituals from bad or
superstitious ones. During the Reformation, for example, Puritans objected to repeated and formulaic
prayers, as well as to any form of magic. Perhaps the
most common concept of ritual language, however, is
a verbal formula of magical or at least pragmatic
the agent of c, that is, at least in typical instances,

to the person who is speaking. The notion that indexicals such as I behave rigidly, even though they are
associated with a descriptive meaning, plays a crucial
role in the classic arguments for the distinction between meaning and content or, in David Kaplans
classic vocabulary, between character and content
(see Kaplan, 1977; Lewis, 1980).
See also: Descriptions, Definite and Indefinite: Philosophical Aspects; Indexicality: Theory; Modal Logic; Possible
Worlds: Philosophical Theories.
Lewis D (1970). Anselm and actuality. Nous 4, 175188.

Kripke S (1980). Naming and necessity (2nd edn.).
Cambridge, MA: Harvard University Press.
Lewis D (1980). Index, context, and content. In Kanger S
hman S (eds.) Philosophy and grammar. Dordrecht:
&O
Reidel.
Marti G (2003). A question of rigidity in new theories of
reference. Nous 37, 161179.
Salmon N (1981). Reference and essence. Princeton, NJ:
Princeton University Press.
Soames S (1998). The modal argument: wide scope and
rigidified descriptions. Nous 32, 122.
Soames S (2002). Beyond rigidity. The unfinished semantic
agenda of naming and necessity. Oxford: Oxford University Press.
Bibliography
Kaplan D (1977). Demonstratives. In Almog J, Perry J &
Wettstein H (eds.) Themes from Kaplan. New York and
Ritual and Religious Language

R A Yelle, University of Toronto, Toronto,
Ontario, Canada
Definition and Varieties

The concept of ritual language presupposes a concept
of ritual, a category beset by numerous definitional
problems. At times it appears that the category of
ritual has become a holding-area for all those forms
of behavior to which we cannot attribute any utilitarian value. This has led both to confusion over the
definition of ritual, and to a common pejorative evaluation of ritual as meaningless, as illustrated by the
modern history of the theorization of this category.
There is a line of influence that extends from Reformation polemics against traditional Catholic liturgical
practices to 19th- and early 20th-century anthropologies characterizing magic as a survival from primitive
culture (Tylor, 1903), as well as a mistaken theory and
praxis outmoded by science (Frazer, 1951). Later sociological interpretations recovered a positive value
for rituals by emphasizing the role that these play in
constructing, reinforcing, and contesting the dynamics of social groups (Durkheim, 1965; Gluckman,
1954; Turner, 1977; Bourdieu, 1977; Lincoln, 1994).
Due to their emphasis on social dynamics, and despite
their frequent application of structuralist analysis to
ritual processes, such interpretations have generally
paid little attention to the specific forms of ritual
language. However, other scholars have emphasized

the performative or pragmatic dimensions of ritual
(Tambiah, 1985), and occasionally conducted sophisticated semiotic analyses of ritual language (Sebeok,
1964; Silverstein and Urban, 1996). Taking all of these
approaches into account produces a more coherent
theory of ritual language.
There exist numerous terms for different types of
ritual language. Each culture of course has its own
terms. Common English terms are spell, charm,
incantation, chant, prayer, and glossolalia.
More recent, theoretical terms include performative
utterance (Austin, 1975), speech act (Searle, 1969),
and indexical icon (Silverstein, 1993, 1998; Silverstein and Urban, 1996). Although by no means exhaustive, this list is representative, and it indicates
some of the important variations among types of
ritual language. Spells and chants are often formulaic
and invariant, whereas glossolalia and certain forms
of prayer are, at least ostensibly, spontaneous and
improvised. Some forms of ritual language aim to
achieve obviously practical objectives, whereas
others, including some forms of prayer, do not. Historically, these very distinctions have informed some
attempts to distinguish good rituals from bad or
superstitious ones. During the Reformation, for example, Puritans objected to repeated and formulaic
prayers, as well as to any form of magic. Perhaps the
most common concept of ritual language, however, is
a verbal formula of magical or at least pragmatic
634 Ritual and Religious Language
effect. Therefore, the problem of how to understand

claims for the efficacy of ritual, including its ability to
bring about real-world objectives through the manipulation of words and other objects, looms large in any
discussion of ritual language. This problem is underscored by the still prevalent pejorative designation of
ritual, especially magic, as mumbo jumbo or hocus
pocus.
Following Redfield (1956), a distinction can be
made between the ritual languages of great or civilizational and little or folk traditions. (This parallels a
similar distinction between official and folk traditions, except that official traditions are not always
great or civilizational.) Of course, both types can
coexist within a single society. The distinction between these types of ritual language can have important consequences in defining social groups. (See
Sociological Dimensions of Ritual Language
below.) There are broader resemblances among the
ritual languages of some of the great traditions that
distinguish them from those of most folk traditions.
The verbal formulas or mantras of the Sanskritic
tradition of Hindu Tantra in South Asia bear more
resemblance to the spells of the Greek Magical
Papyri, or to Kabbalistic formulas, than they do to
Bengali folk mantras, despite greater geographical
and cultural proximity to the latter. The main distinction from folk spells consists in the development of a
more elaborate code, represented by increasingly intricate linguistic patterns, as exemplified by the extended palindromes of some Greek and Sanskrit
spells. Such devices, which are difficult or impossible
to recognize in an oral tradition, may reflect in part
an increased reliance on writing, as well as the presence of esoteric traditions necessary for the elaboration of a code, although this last feature is of course
also present in many oral traditions.
The distinction between oral cultures and cultures
in which writing or printing has become dominant as
a mode of communication is of great importance for
ritual language. There is some evidence that the critique and decline of certain forms of ritual language
may be associated with writing. The importance of
poetic formulas as mnemonic devices and aids to
performance in oral cultures has been recognized in
recent decades (Lord, 1960). It has been argued further that writing not only replaces poetry as a technique of cultural transmission, but may even lead to a
new, skeptical episteme critical of the irrationality
of poetry (Havelock, 1963; Goody, 1977). The connection between poetry and ritual or magical language is sometimes reinforced in these critiques,
making these issues relevant to an account of ritual
language. (See Historical Dimensions of Ritual
Language below.)
Nonritual Religious Language

Although the focus of this article is primarily on ritual
language, other types of religious language are
properly described here in relation to ritual language.
In religion as in every other domain of human culture,
language plays a central role as the most important
means of communication, a fact recognized in many
religious traditions. Some well-known examples
are the creation of the world through the power of
Gods speech in the Hebrew Book of Genesis, the
identification of Jesus as the Word of God in the
Christian Gospel of John, and the various cosmogonies developed in Hinduism from Vedic times onward
in which speech is the agent or instrument of
creation (Padoux, 1990). Many religious traditions
acknowledge the power of the word by the formation
of canons (Smith, 1982), especially of the written
variety called scripture. Although canon and codification are broader phenomena that have also been
associated with modernity and colonialism, the
specifically religious conflict between the Peoples of
the Book and the idolaters had special significance
for the historical development of ritual languages.
(See Historical Dimensions of Ritual Language
below.)
Myth is another subcategory of religious language
that has received much attention from linguists and
anthropologists, especially following Le vi-Strausss
application of structuralist linguistic theories to
myth. The fact that myths are a form of language,
whereas ritual comprises a range of behaviors not
limited to the verbal, may explain why structuralist
analyses of ritual came later than those of myth.
The identification of myth and ritual as fundamental
categories of religion has precipitated a number of
theories concerning the relation between these categories. The so-called myth and ritual school developed the idea that myth derives from and attempts to
explain a preexisting ritual complex. On the other
hand, Eliade (1954) argued that myth describes the
actions of the gods or ancestors at the beginning of
time, and that ritual refers to myth as the archetype
for human actions in the present. As illuminating as
these interpretations occasionally are, it is important
to note that many rituals are largely independent
from myth, and vice versa.
Structure and Function of Ritual

Language
For purposes of analysis, it is useful to separate the
structural or formal aspects of ritual language from
their functional aspects or pragmatic effects, although
the goal is to arrive at a better understanding of the
contribution of form to function. The formal differences of ritual from ordinary language were described in great detail by Malinowski in his study of
Kiriwinian (Trobriand Islands) spells (Malinowski,
1935). These spells exhibited a coefficient of weirdness that came from their use of archaisms and
apparently meaningless forms, onomatopoeias, repetitive and rhythmic chanting, extensive metaphors,
and allusions to myth. Malinowski argued that
many magic words were meaningless, but also that
the various stylistic features of spells somehow supported the creative metaphor of magic, or the belief
that ritual is able to bring about its intended result.
Working in part from the same data, Tambiah (1985)
argued for the persuasive or performative function of
a number of these stylistic features. However, his
model emphasized the semantic figures, especially
the metaphors and metonyms, of ritual language,
rather than its rhythmic form. It is this latter feature
that raises what might be called the abracadabra
problem: how do we explain the prevalence in ritual
language of so many apparently nonsensical, repetitive sounds, as evidenced not only in numerous ritual
traditions (e.g., the Ephesian Words beginning aski
kataski. . . found in many Greek spells, and the Hindu
Tantric seed (bja) mantras such as hrm klm srm),
but also in pejorative terms such as mumbo jumbo?
The dismissal of such forms as meaningless (e.g., by
Staal, 1996) does not account for their frequency of
occurrence in ritual language. For any particular term
such as abracadabra or hocus pocus, one may be
able to provide a probable etymon of which the magical word is a corruption. However, this does not
explain how, as actually used in ritual discourse,
such magic words function.
One possible explanation for such stylistic features
is that they serve primarily to differentiate ritual from
ordinary language. This hypothesis accords with one
of the fundamental insights of linguistic structuralism: namely, that any distinction in the value of a
linguistic sign depends upon an associated distinction
in its form. The principle of binary opposition discovered by Saussure may also serve at the level of genres
to distinguish ordinary language, on the one hand,
from poetry, literature, or ritual discourse, on the
other. Indeed, the thesis that poetry represents simply
a deviation from ordinary language has been articulated within structuralist poetics. This thesis is
paralleled within ritual studies by Bells emphasis on
techniques or strategies of ritualization that serve to
differentiate ritual from ordinary discourse or behavior (Bell, 1992). These theories may account for the
fact that many forms of ritual language are incomprehensible to their hearers and even, on occasion,
their speakers. In such cases, it is not any semantic
content, but rather the formal differences of the ritual

medium that constitute its message. The preservation
of the mass in Latin long after that language had
ceased to be a vernacular, and the dissemination of
Sanskrit mantras in transliterated rather than translated form as far as Japan, reflect a more basic strategy to distinguish ritual from the mundane or profane,
as well as to conserve the prestige of origins, and to
create mystery.
An objection to such theories is that they fail to
explain the prevalence of poetic repetition and other
particular devices, including metaphor and metonymy, in the ritual languages of different cultures. If
ritual were only a matter of constructing difference,
then any deviation from ordinary language would do,
and we should expect randomness in the variety and
distribution of stylistic features in ritual language.
What we often find instead is an extreme form of
the poetic function as defined by Jakobson (1960): a
projection of equivalence into the sequence of language, manifested as versification, rhyme, semantic
repetition, etc. These devices are especially prevalent
in magic. What explains this phenomenon?
Silverstein has shown that ritual performance often
constitutes an indexical icon, in which forms of repetition or resemblance in the text of a ritual increasingly
point to their context. This is similar to the mechanism
of (nonverbal) sympathetic magic, which, as theorists
from Tylor and Frazer onward have noted, interprets a
relation of similarity (metaphor, icon) or contiguity
(metonym, index) between the ritual and its intended
result as a causal (indexical) relation (cf. Leach, 1976).
Just as repeatedly poking a pin into the left arm of a
voodoo doll is supposed to bring about a corresponding
injury in ones intended victim, or pouring water on the
ground is supposed to bring rain from the sky, the
poetic imitation in ritual language is supposed to
point beyond itself to bring about a real-world result.
In the case of many folk charms, the accumulation of
rhythmic sound appears to be the mechanism by which
ritual language increasingly signals its intended result.
An example is Rain, rain, go away/Come again some
other day. Resemblances (icons) on the phonetic level
of the text of the ritual reinforce the ritual as a signal
(index) of a goal in its context: in this case, the cessation
of rain (cf. Sebeok, 1964).
As compared with such folk charms, the mechanism
of Hindu Tantric mantras is more complex. Mantras
often are made effective (siddha) by being converted
into palindromic diagrams of several processes of creation, including the cosmogony conceived as a cycle
of evolution and involution, and sexual reproduction
(Yelle, 2003). Ritual performance of the mantra both
imitates and instantiates the creative power of speech
as the source of the cosmos. Syllables of special power
called seeds (bja) surround the central portion of the

mantra, which often declares its real-world goal, such
as subjugating an enemy. By verbally tracing the sequence of creation, the mantra leads up to and indexes this real-world goal, symbolically bringing it into
existence.
Such examples suggest that at least some types of
ritual language do not operate by reinforcing a difference between ritual and ordinary language, but rather
by attempting to overcome any difference or gap between ritual language and its context of use. Ritual
tries to construct a language that is directly connected
to, and even capable of influencing, reality. This hypothesis appears to account for two additional devices
found in ritual language. Many spells enumerate an
entire paradigm class, which may be a set of objects
such as body parts (cf. Tambiah, 1985), but also a set
of rhyming words or, in the case of Tantric mantras
and Greek spells, all or a portion of the sound system
or alphabet. Such extreme forms of repetition augment ritual language as an indexical sign of its goal.
This pointing function is further illustrated by socalled counting-out spells which, by counting from
ten down to zero (or, in the case of some Greek spells,
removing all of the letters of the spell one by one),
signal the end and accomplishment of the spell.
These examples suggest a different evaluation of
the relation between emic (culture-internal) and
etic (comparatively cross-cultural) theories of performance or pragmatic function. From an emic perspective, the forms of poetic imitation employed in
ritual language appear to construct a language that
bears an especially close indexical relation to nature.
However, from an etic or outsiders perspective, such
devices serve primarily to construct a persuasive illusion of a perfect and effective language.
These considerations require us to reevaluate what
is meant by ritual performance. In recent decades,
following Austins definition of a performative utterance and Searles refinement of this with the notion of
a speech act, the theory that ritual is a form of performance has gained broader currency (Tambiah, 1985).
Austins original concept has been applied to many
cases, including magic, to which it did not on its own
terms claim to extend. This is of course not a reason
for automatically dismissing such applications. However, it is a reason for revisiting the concept of a
performative, especially as applied to ritual.
Austin recognized that many utterances accomplish
something through the simple fact of their being
uttered. Statements such as I christen this ship or I
thee wed do not merely say something, but do something. Linguists would agree that such statements have
pragmatic and not merely semantic value. To this extent, Austins concept of a performative represents a
valuable corrective against the semantic bias of much

modern philosophy, with its emphasis on propositional
content. As we have seen, a similar semantic bias also
informs many interpretations of ritual language as
meaningless.
However, there are a few problems with Austins
theory. First, his concept of performance or pragmatic function is too restrictive. If the conditions of
its use are fulfilled, a performative utterance accomplishes something through the mere act of utterance:
mutual promises may themselves constitute a valid
legal contract. On the other hand, many rituals, especially of the magical variety, point to some real-world
goal, such as the beginning or cessation of rain, something that may or may not occur automatically
upon the completion of the ritual. The fiction may
be that ritual achieves these goals ex opere operato,
but the reality is otherwise. The concept of performativity or pragmatic function must be broadened, if
applied to such cases, to allow that ritual sometimes performs by constructing a persuasive illusion
of efficacy.
A second problem with Austins theory is that he
omitted any account of the contribution of poetry to
ritual performance. In his view, performatives, like
other ceremonial doings, were purely conventional
in form, formulaic. On the contrary, we have already
seen that many types of ritual language employ poetry
in order to augment if not to bring about the performative or pragmatic function of ritual. Some historical roots of Austins neglect of the poetry of ritual are
examined below.
Sociological Dimensions of Ritual

Language
Ritual languages may not be distributed evenly within
a society. Like other ritual techniques, they are often
the property of particular social groups. Individuals
who are not members of these groups may be denied
training in the techniques of ritual performance, or
access to the cultural code that permits the full comprehension of the ritual, or even the opportunity to
observe a ritual performance. Esotericism results
whenever there is a deliberate and sufficiently rigorous restriction of access to ritual to an in-group of
initiates. Public rituals, of course, may also reinforce
divisions within the social order. Proprietary rituals
have a role in defining social groups, for example in
male and female ritual societies. In some cases, the
rituals with the greatest cultural prestige those of
the official tradition(s) are reserved for male practitioners, and women may develop expertise in other,
folk rituals, such as trance, mediumship, possession,
and witchcraft, which are understood to exist at a less
important level of the social structure. In Catholicism

the performance of the sacraments is reserved for
ordained males. In Hindu Tantra, there are numerous
prohibitions against the use of the sacred syllable om
by women or servants (s u dras, the lowest division of
the social order), who are instead enjoined to substitute some other syllable. Given the symbolic significance of om as the primordial utterance of the first
man, which ushered in creation, these ritual injunctions also encode the idea that women and lowerclass individuals should be denied access to sacred,
effective speech: they are made (ritually) voiceless.
Sometimes ritual languages may be used not to
reinforce the social order, but to contest ones position
within, or criticize, or even overthrow that order.
Examples are Radcliffe-Browns joking relationships
and Gluckmans rituals of rebellion (Radcliffe-Brown,
1952; Gluckman, 1954). Such rituals are often just a
way of blowing off a little steam. Moreover, Bloch
(1989) pointed out that the more stereotyped, formalized, and scripted ritual discourse becomes, the more
it tends to reinforce conventional authority, by rendering deviation from the script in performance less
likely. However, even rigidly scripted rituals may get
out of hand. Lincolns account of the discursive construction of authority allows for just such occasions
of corrosive discourse. These accounts raise difficult
questions of agency, or of the relation between individuals and social structures. Bourdieus concept of
habitus as the flexible cultural structures that both
condition individual responses, and are susceptible
to strategic manipulation by individuals, seeks to
address this fundamental issue.
Questions of agency and intentionality are also
of paramount importance in connection with such
rituals as trance, mediumship, and possession, where
there is a displacement or denial of the ritual practitioners agency. Perhaps the most well-known example of ritual language of this sort is glossolalia, in
which the Holy Spirit speaks through the practitioner
in what is understood to be an unscripted performance. Even these spontaneous forms, however,
are learned social behaviors that exhibit phonological
constraints reflecting the natural languages of their
users (Samarin, 1972; Goodman, 1972).
Relation between Verbal and Other

Modes of Ritual
Earlier tendencies to focus on myth, due its linguistic
medium and consequent assumed meaningfulness, to
the relative neglect of certain nonverbal forms of
ritual are now in the first stages of reversal. More
attention is now being given to such media as dance.
However, these studies have to contend with a basic
defect of ethnographic method: the lack of any system

of notation of bodily movement that is both accurate
and widely used. This parallels the problem of recording and interpreting the special modes of pronunciation or chanting of ritual language. Although tape
recordings can capture such forms in the raw, there
is still a significant difficulty in the recognition
and interpretation of their distinctive features from
analytic study of notated transcripts.
Certain stylistic features, such as imitation, appear
to transcend the divide between the verbal and nonverbal modes of ritual. Aristotle noted that the modes
of imitation (mimesis) are different in different arts:
in music, rhythm and harmony would be the modes
employed (Poetics 1.1), whereas in painting, color
and line would presumably be the relevant modes.
The concept of mimesis itself has been traced to the
ritual dances that formed the basis of Greek tragedy
(Koller, 1954), which suggests that the original forms
of imitation were not verbal at all, but gestural. The
use of imitation in the nonverbal dimensions of ritual
was documented already in Frazers account of sympathetic magic. As previously described, ritual language often uses poetic imitation in a similar way.
The Hindu Tantric womb sign (yonimudra ), which
iconically depicts the embrace of the womb, is a
gestural equivalent and substitute for the verbal
enveloping of the mantra with magic syllables. Both
types of icon diagram basic processes of creation,
including sexual intercourse. In such cases, verbal
and nonverbal forms of ritual behavior may carry
an identical meaning and pragmatic function. The
fact that certain semiotic patterns transcend the divide between language and movement reinforces the
conclusion that movement shares with language a
cultural and cognitive basis.
Ritual frequently coordinates verbal and nonverbal
behaviors through superimposition, simultaneity, and
sequencing. The entextualization of the ritual performance (cf. Silverstein and Urban, 1996) constructs a
diagram or diagrammatic icon in which speech and
movement are fused and reciprocally imitative. For
example, in Hindu Tantra, the syllables of the mantra
may be laid out on the body of a ritual practitioner,
or visualized as moving back and forth within the
body, as a representation of the processes of creation
otherwise verbally diagrammed by the mantra when
spoken.
If the same functions may be served alternatively by
either verbal or nonverbal modes of ritual, then why
should the coordination of these modes be so frequent? One of the characteristics of ritual is that it
builds in redundancies that attract attention or otherwise heighten the communicative force of its message.
These redundancies may occur not only within but
across semiotic modes. Bloch (1989) has suggested

further that the accumulation of symbols, song,
dance, and features of articulation within ritual
discourse is a way of enhancing its formality and
consequent reinforcement of traditional authority.
Historical Dimensions of Ritual

Language
The historical or diachronic dimensions of ritual language have received less systematic consideration
than its structural or synchronic dimensions. This is
in part due to the emphasis within modern linguistics
on the study of language as a holistic system. Saussure
famously drew a distinction between the system of a
language (langue) and its instantiation in speech
events (parole). Yet whether modern students of language choose to focus on the former or, especially
through real-time discourse analysis, the latter, the
end result is the same: historical transformations in
linguistic practices and emic theories of language are
largely left out of account.
This neglect presents a particular problem for the
understanding of ritual language. One important historical development in linguistic practices is a diminution, within certain domains of certain cultures, of the
poetic forms and pragmatic functions associated with
ritual. We must be careful here to replicate neither the
nostalgic associations of pre-modern folk culture with
poetry promulgated by some Romantics, nor the triumphal evolutionary accounts of the replacement of
magic by science advanced by early anthropologists.
However, there is a great deal of evidence that, at
some point in the past, the nexus of poetry and
magic was more prominent than it is now. Grimms
study of poetic parallelisms in early Germanic law
affords some examples (Grimm, 1957). Rhythmic formulas such as unbidden and unbought, or to have
and to hold, were common in early English law as
well, where they contributed to the binding force of
many vows, oaths, and declarations. What happened
to such formulas and the mode of ritual that they
represented? One answer is that post-Enlightenment
modernity has emphasized literalism, or plain speaking and the semantic content of language, at the expense of both poetry and magic. Poetry has been
reduced to a mode of entertainment, and few people
believe in what goes for magic these days. The semantic bias has reinforced the view of ritual as meaningless. However, these are only observations, not
explanations.
One explanation that attempts to account for the
perceived decline in formalized ritual is Tambiahs
concept of ritual involution. The very features
that contribute to the performative aspects of ritual,
such as increased formalization, over time may lead

to a rejection of that ritual as rigid and empty. Certainly both attitudes toward ritual formalization
positive and negativemay be observed. However,
Tambiahs concept provides no diagnosis of the conditions that would result in one attitude rather than
the other, diametrically opposed attitude being
adopted. Moreover, he has framed ritual involution
largely as a subjective rather than as an historical
process, despite the fact that the pejorative evaluation
of ritual as stereotyped and meaningless is especially
prominent in modernity, and must therefore be given
an historical explanation.
Another explanation, mentioned at the outset of
this article, builds from recent studies on the effects
of the transition from an oral culture to one based on
writing (Havelock, 1963; Goody, 1977). According
to what we may call the literacy hypothesis, poetry
was necessary in oral culture as a mnemonic device.
Writing not only replaced this function of poetry, but
introduced an emphasis on literal meaning and a
skepticism regarding certain modes of rhetoric. However, such theories do not fully explain why the obsolescence of poetry as a mnemonic technique should
have led to such pejorative evaluations. Nor do they
account for the fact that, in the prime example addressed by the literacy hypothesis, that of ancient
Greece, the focus of such criticisms was on the persuasive rather than the mnemonic function of poetry.
The persuasive or rhetorical function of poetry established its association with magic for both the sophists
and their opponents, the philosophers (see Gorgias,
Helen; Plato, Republic, Book 10). As we saw, it is this
crucial nexus of poetic form and persuasive function
that characterizes many rituals.
Apart from ancient Athens, another important moment in the development of ritual language occurred
during the Reformation. The Protestant emphasis on
the literal interpretation of a canon of scripture, and
on translation into the vernacular, led to the demise of
certain older liturgical forms. This represented not
merely a modern rationalization of religion, but also
a recrudescence of certain ancient Jewish and Christian religious ideas, including especially the prohibition against idolatry. Christ prohibited the use of
vain repetitions in prayer (Matt. 6:7, Authorized
Version [1611]). Some Puritans associated this prohibited practice with the worship of idols, especially
for magical purposes, and with attempts to persuade the deity rhetorically (e.g., Calvins Institutes;
Matthew Henrys Commentary). They applied this
prohibition against repetitive or formalized prayers
employed by Catholics and also by the established
Church of England. Thomas Cranmer, the author of
the Book of common prayer (1549), explained in his
introduction to that work that he sought to eliminate

many superstitious rituals, including vain repetitions, from the liturgy. Although he did not go as
far as some Puritans would have liked, one consequence of his effort was that more poetic versions of
the marriage oath disappeared. For example, the
wifes promise to be bonour and buxom at bed and
at board was replaced by the now familiar to love,
cherish, and to obey. The sole remaining rhythmic
(as opposed to semantic) parallelism was the declaration to have and to hold, an old formula of ownership attested in both English and German law.
Although the critique of vain repetitions had not
been applied directly against such rhythmic formulas
in the marriage vow, it contributed to a literalism that
ultimately caused the demise of these formulas, as
well as of Latin as a ritual language. The irony is
that Austins own prime example of a performativethe declaration I do (sic) during the marriage
ceremonyunwittingly continued this repression of
the poetry of ritual, by failing to acknowledge the
historical context and development of the marriage
oath.
As in ancient Greece, the literacy hypothesis appears
to account for some of these developments. It has been
suggested that the poetic devices of early English law
served a mnemonic function in an oral culture. The
Reformation, leading to the Book of common prayer,
reflected not the discovery of alphabetic writing as in
ancient Greece, but instead a culture of printing based
on the technological innovation of movable type in
15th-century Europe. The result in each case was similar: an increase in literacy, and in literalism. However,
in each case these developments were facilitated not
only by writing, but also by polemics against the nexus
of poetic form and persuasive function constituted
in ritual.
It is important to recognize that religion stands on
both sides of this debate over ritual language, both as
the matrix for traditional rituals and as the source of
certain criticisms of these rituals as idolatry. A survey
of the development of linguistic empiricism in Britain
reinforces this conclusion. Following Francis Bacons
(15611626) attack on idols of the marketplace,
John Locke (16321704) excoriated the habit of taking words for things, or inferring from the existence
of a name to the existence of the thing named. Jeremy
Benthams (17481832) critique of such fictions
drew explicitly on religious prohibitions against idolatry. Among the fictions he sought to proscribe were
ritualized oaths used in the courtroom. As late as
the second half of the 19th century, Friedrich Max
Mu llers (18231900) theory of myth as a disease
of language still reflected such criticisms of verbal
idolatry. Nor were these developments confined to
England. In colonial India, the gospel verse against

vain repetitions was used to proscribe the repeated chanting of mantras. These polemics influenced
the religious literalism of such Hindu Reformers as
Rammohun Roy (1772?1833) and Dayananda
Sarasvati (18241883), who championed the canon
of the Vedas against such forms of idolatrous ritual.
In conclusion, a brief historical survey of certain
developments regarding ritual language underscores
the need to consult both historical and anthropological data when proposing a theory of ritual language.
As linguists, we must be concerned not only with the
elaboration of a neutral science of the laws of language, but also with the legislations of language made
by particular cultures at particular times, that determine which among the possible forms and functions
of language will be permitted or prohibited.
See also: Glossolalia; Literacy Practices in Sociocultural
Perspective; Performance in Culture; Religion and Literacy; Ritual and Religious Language.
Bibliography
Alper H (ed.) (1989). Mantra. Albany: SUNY Press.
Austin J L (1975). How to do things with words.
Cambridge: Harvard University Press.
Bell C (1992). Ritual theory, ritual practice. New York:
Oxford University Press.
Betz H D (ed.) (1992). The Greek magical papyri in
translation (2 vols) (2nd edn.). Chicago: University of
Chicago Press.
Bloch M (1989). Ritual, history, and power: selected papers
in anthropology. London: Athlone Press.
Bourdieu P (1977). Outline of a theory of practice. Nice R
(trans.). Cambridge: Cambridge University Press.
Durkheim E (1965). The elementary forms of the religious
life. Swain J W (trans.). New York: The Free Press.
Eliade M (1954). The myth of the eternal return: or, cosmos
and history. Trask W (trans.). Princeton: Princeton
University Press.
Frazer J G (1951). The golden bough: a study in magic and
religion (1 vol. abridged edn.). New York: Macmillan.
Gluckman M (1954). Rituals of rebellion in south-east
Africa. Manchester: Manchester University Press.
Goodman F (1972). Speaking in tongues: a cross-cultural
study of glossolalia. Chicago: University of Chicago
Press.
Goody J (1977). The domestication of the savage mind.
Cambridge: Cambridge University Press.
Grimm J (1957). Von der poesie im recht (reprint edn.).
Darmstadt: Hermann Gentner Verlag.
Havelock E A (1963). Preface to Plato. Cambridge:
Harvard University Press.
Jakobson R (1960). Closing statement: linguistics and poetics. In Sebeok T (ed.) Style in language. Cambridge:
MIT Press. 350377.

Koller H (1954). Die mimesis in der antike. Bern: A.
Francke.
Leach E (1976). Culture and communication. Cambridge:
Cambridge University Press.
Lincoln B (1994). Authority: construction and corrosion.
Chicago: University of Chicago Press.
Lord A (1960). The singer of tales. Cambridge: Harvard
University Press.
Malinowski B (1935). Coral gardens and their magic (2
vols). New York: American Book Company.
Padoux A (1990). Va c: the concept of the word in selected
Hindu tantras. Gontier J (trans.). Albany: SUNY Press.
Radcliffe-Brown A R (1952). Structure and function in
primitive society: essays and addresses. London: Cohen
& West.
Redfield R (1956). Peasant society and culture. Chicago:
University of Chicago Press.
Samarin W (1972). Tongues of men and angels: the religious language of Pentecostalism. New York: Macmillan.
Searle J (1969). Speech acts: an essay in the philosophy of
language. Cambridge: Cambridge University Press.
Sebeok T (1964). The structure and content of Cheremis
charms. In Hymes D (ed.) Language in culture and
society. New York: Harper and Row. 356371.
Ritwan
Silverstein M (1993). Metapragmatic discourse and

metapragmatic function. In Lucy J A (ed.) Reflexive
language: reported speech and metapragmatics. Cambridge: Cambridge University Press. 3358.
Silverstein M (1998). The improvisational performance of
culture in realtime discursive practice. In Sawyer R K
(ed.) Creativity in performance. Greenwich, CT: Ablex
Publishing Corp. 265311.
Silverstein M & Urban G (eds.) (1996). Natural histories of
discourse. Chicago: University of Chicago Press.
Smith J Z (1982). Imagining religion: from Babylon to
Jonestown. Chicago: University of Chicago Press.
Staal F (1996). Ritual and mantras: rules without meaning.
Delhi: Motilal Banarsidass.
Tambiah S (1985). Culture, thought, and social action.
Turner V (1977). The ritual process: structure and antistructure. Ithaca: Cornell University Press.
Tylor E B (1903). Primitive culture (4th rev. edn.). London:
John Murray.
Yelle R A (2003). Explaining mantras: ritual, rhetoric, and
the dream of a natural language in Hindu tantra. London:
Routledge.
See: Algonquian and Ritwan Languages.
Robins, Robert Henry (19212000)

A Linn, University of Sheffield, Sheffield, UK
Robert Henry Robins, known as Bobby to his friends

and colleagues, is best remembered for his A short
history of linguistics, which was originally published
in 1967 and entered its 4th and final edition in 1997.
Robins was a great advocate for the study and teaching of the history of linguistics, and while he contributed to research in a number of areas of linguistics, it
was the study of the history of the discipline that
formed the backbone to his lifes work. His 1951
book Ancient and medieval grammatical theory in
Europe with particular reference to modern linguistic
doctrine grew out of a series of lectures he was persuaded to give by J. R. Firth at Birkbeck College,
London, in 1950. Later historiographical work extended to studies of the London School, of which he
was himself a part, the 16th-century grammarian
of English William Bullokar, the 18th-century philosopher of language, E tienne Bonnot de Condillac,
the 18th-century orientalist Sir William Jones, and
the 19th-century Neogrammarians, as well as to
reflections on the nature of the subject itself. In his
last book (Robins, 1993) he returned to the classical
interests with which he started.
His research energies were not limited to the classical languages and the history of linguistics, however.
In 1951 he worked on the Californian language Yurok
and published on this language throughout his career
as well as on the Javan language Sundanese (Sunda).
His interest in endangered languages resulted in the
book Endangered languages (Robins and Uhlenbeck,
1991), published under the auspices of the Comite
International Permanent de Linguistes, of which he
was president. Alongside the Short history, the other
work with which Robinss name is particularly associated, is his General linguistics: an introductory survey (1964; 4th edn., 1989). As lecturer, then reader,

Koller H (1954). Die mimesis in der antike. Bern: A.
Francke.
Leach E (1976). Culture and communication. Cambridge:
Lincoln B (1994). Authority: construction and corrosion.
Chicago: University of Chicago Press.
Lord A (1960). The singer of tales. Cambridge: Harvard
University Press.
Malinowski B (1935). Coral gardens and their magic (2
vols). New York: American Book Company.
Padoux A (1990). Vac: the concept of the word in selected
Hindu tantras. Gontier J (trans.). Albany: SUNY Press.
Radcliffe-Brown A R (1952). Structure and function in
primitive society: essays and addresses. London: Cohen
& West.
Redfield R (1956). Peasant society and culture. Chicago:
University of Chicago Press.
Samarin W (1972). Tongues of men and angels: the religious language of Pentecostalism. New York: Macmillan.
Searle J (1969). Speech acts: an essay in the philosophy of
language. Cambridge: Cambridge University Press.
Sebeok T (1964). The structure and content of Cheremis
charms. In Hymes D (ed.) Language in culture and
society. New York: Harper and Row. 356371.
Ritwan
Silverstein M (1993). Metapragmatic discourse and

metapragmatic function. In Lucy J A (ed.) Reflexive
language: reported speech and metapragmatics. Cambridge: Cambridge University Press. 3358.
Silverstein M (1998). The improvisational performance of
culture in realtime discursive practice. In Sawyer R K
(ed.) Creativity in performance. Greenwich, CT: Ablex
Publishing Corp. 265311.
Silverstein M & Urban G (eds.) (1996). Natural histories of
discourse. Chicago: University of Chicago Press.
Smith J Z (1982). Imagining religion: from Babylon to
Jonestown. Chicago: University of Chicago Press.
Staal F (1996). Ritual and mantras: rules without meaning.
Delhi: Motilal Banarsidass.
Tambiah S (1985). Culture, thought, and social action.
Turner V (1977). The ritual process: structure and antistructure. Ithaca: Cornell University Press.
Tylor E B (1903). Primitive culture (4th rev. edn.). London:
John Murray.
Yelle R A (2003). Explaining mantras: ritual, rhetoric, and
the dream of a natural language in Hindu tantra. London:
Routledge.
See: Algonquian and Ritwan Languages.
Robins, Robert Henry (19212000)

A R Linn, University of Sheffield, Sheffield, UK
Robert Henry Robins, known as Bobby to his friends

and colleagues, is best remembered for his A short
history of linguistics, which was originally published
in 1967 and entered its 4th and final edition in 1997.
Robins was a great advocate for the study and teaching of the history of linguistics, and while he contributed to research in a number of areas of linguistics, it
was the study of the history of the discipline that
formed the backbone to his lifes work. His 1951
book Ancient and medieval grammatical theory in
Europe with particular reference to modern linguistic
doctrine grew out of a series of lectures he was persuaded to give by J. R. Firth at Birkbeck College,
London, in 1950. Later historiographical work extended to studies of the London School, of which he
was himself a part, the 16th-century grammarian
of English William Bullokar, the 18th-century philosopher of language, Etienne Bonnot de Condillac,
the 18th-century orientalist Sir William Jones, and
the 19th-century Neogrammarians, as well as to
reflections on the nature of the subject itself. In his
last book (Robins, 1993) he returned to the classical
interests with which he started.
His research energies were not limited to the classical languages and the history of linguistics, however.
In 1951 he worked on the Californian language Yurok
and published on this language throughout his career
as well as on the Javan language Sundanese (Sunda).
His interest in endangered languages resulted in the
book Endangered languages (Robins and Uhlenbeck,
1991), published under the auspices of the Comite
International Permanent de Linguistes, of which he
was president. Alongside the Short history, the other
work with which Robinss name is particularly associated, is his General linguistics: an introductory survey (1964; 4th edn., 1989). As lecturer, then reader,
Role and Reference Grammar 641
then professor at the School of Oriental and African

Studies in London, Robins was a committed teacher,
and he continued to teach at the Universities of
Cambridge and Luton even after his retirement.
Collections of his papers are to be found in Robins
(1970, 1998). An illuminating interview with Robins
is in Orbis Supplementa 9, 6188, an autobiographical account in Brown and Law (2002: 249261), and
a biography and full bibliography in Law and Hu llen
(1996: 942). By his own admission, the two features
that characterize Robinss work, whatever he was
working on, are the study of language or of linguistics
in its context and the avoidance of a dogmatic position. As a person, Robins was an old-fashioned English gentleman and a representative of an era now
only fondly remembered.
See also: Bullokar, William (c. 15311609); Condillac, Etienne Bonnot de (17141780); Firth and the London School;
Firth, John Rupert (18901960); Jones, William, Sir (1746
1794); Modern Linguistics: 1800 to the Present Day; Neogrammarians; Western Linguistic Thought Before 1800.
Bibliography
Brown K & Law V (eds.) (2002). Linguistics in Britain:
personal histories. Oxford: Blackwell.
Law V & Hu llen W (eds.) (1996). Linguists and their
diversions: a festschrift for R. H. Robins on his 75th
birthday. Mu nster: Nodus Publikationen.
Robins R H (1970). Diversions of Bloomsbury: selected
writings on linguistics. Amsterdam/London: NorthHolland Publishing Company.
Robins R H (1993). The Byzantine grammarians: their
place in history. Berlin/New York: Mouton de Gruyter.
Robins R H (1998). Texts and contexts: selected papers on
the history of linguistics. Law V (ed.). Munster: Nodus
Publikationen.
Robins R H & Uhlenbeck E M (1991). Endangered languages. Oxford/New York: Berg.
Role and Reference Grammar

R D Van Valin, Jr., University at Buffalo, The State
University of New York, Buffalo, NY, USA
Role and Reference Grammar (RRG) (Van Valin,

1993a, 2005; Van Valin and LaPolla, 1997) grew
out of an attempt to answer two basic questions:
(i) what would linguistic theory look like if it were
based on the analysis of languages with diverse structures such as Lakhota (Lakota), Tagalog, and Dyirbal, rather than on the analysis of English?, and (ii)
how can the interaction of syntax, semantics and
pragmatics in different grammatical systems best be
captured and explained? RRG takes language to be a
system of communicative social action, and accordingly, analyzing the communicative functions of
grammatical structures plays a vital role in grammatical description and theory from this perspective. It is
a monostratal theory, positing only one level of syntactic representation, the actual form of the sentence.
The overall organization of the theory is given in
Figure 1.
syntax), because they are not universal and hence

necessarily impose aspects of structure on at least
some languages where it is not appropriate. The
RRG conception of clause structure, the layered
structure of the clause (LSC), is made up of the nucleus, which contains the predicate(s), the core, which
contains the nucleus plus the arguments of the predicate(s), and the clause. Syntactic arguments occurring in the core are referred to as core arguments, and
they may be direct or oblique. Direct core arguments
are those not marked by an adposition, in languages
such as English and German, or those marked by
direct cases (nominative, accusative, dative or ergative, absolutive, dative) in case-marking languages
like Russian or Dyirbal. Oblique core arguments
are marked by adpositions or oblique cases, e.g.,
instrumental and locative. Modifying each level of
Central Concepts of the Theory

Clause Structure
RRG rejects the standard formats for representing clause structure (grammatical relations, X-bar
Figure 1 Organization of role and reference grammar.
then professor at the School of Oriental and African

Studies in London, Robins was a committed teacher,
and he continued to teach at the Universities of
Cambridge and Luton even after his retirement.
Collections of his papers are to be found in Robins
(1970, 1998). An illuminating interview with Robins
is in Orbis Supplementa 9, 6188, an autobiographical account in Brown and Law (2002: 249261), and
a biography and full bibliography in Law and Hullen
(1996: 942). By his own admission, the two features
that characterize Robinss work, whatever he was
working on, are the study of language or of linguistics
in its context and the avoidance of a dogmatic position. As a person, Robins was an old-fashioned English gentleman and a representative of an era now
only fondly remembered.
See also: Bullokar, William (c. 15311609); Condillac, Etienne Bonnot de (17141780); Firth and the London School;
Firth, John Rupert (18901960); Jones, William, Sir (1746
1794); Modern Linguistics: 1800 to the Present Day; Neogrammarians; Western Linguistic Thought Before 1800.
Bibliography
Brown K & Law V (eds.) (2002). Linguistics in Britain:
personal histories. Oxford: Blackwell.
Law V & Hullen W (eds.) (1996). Linguists and their
diversions: a festschrift for R. H. Robins on his 75th
birthday. Munster: Nodus Publikationen.
Robins R H (1970). Diversions of Bloomsbury: selected
writings on linguistics. Amsterdam/London: NorthHolland Publishing Company.
Robins R H (1993). The Byzantine grammarians: their
place in history. Berlin/New York: Mouton de Gruyter.
Robins R H (1998). Texts and contexts: selected papers on
the history of linguistics. Law V (ed.). Munster: Nodus
Publikationen.
Robins R H & Uhlenbeck E M (1991). Endangered languages. Oxford/New York: Berg.
Role and Reference Grammar

R D Van Valin, Jr., University at Buffalo, The State
University of New York, Buffalo, NY, USA
Role and Reference Grammar (RRG) (Van Valin,

1993a, 2005; Van Valin and LaPolla, 1997) grew
out of an attempt to answer two basic questions:
(i) what would linguistic theory look like if it were
based on the analysis of languages with diverse structures such as Lakhota (Lakota), Tagalog, and Dyirbal, rather than on the analysis of English?, and (ii)
how can the interaction of syntax, semantics and
pragmatics in different grammatical systems best be
captured and explained? RRG takes language to be a
system of communicative social action, and accordingly, analyzing the communicative functions of
grammatical structures plays a vital role in grammatical description and theory from this perspective. It is
a monostratal theory, positing only one level of syntactic representation, the actual form of the sentence.
The overall organization of the theory is given in
Figure 1.
syntax), because they are not universal and hence

necessarily impose aspects of structure on at least
some languages where it is not appropriate. The
RRG conception of clause structure, the layered
structure of the clause (LSC), is made up of the nucleus, which contains the predicate(s), the core, which
contains the nucleus plus the arguments of the predicate(s), and the clause. Syntactic arguments occurring in the core are referred to as core arguments, and
they may be direct or oblique. Direct core arguments
are those not marked by an adposition, in languages
such as English and German, or those marked by
direct cases (nominative, accusative, dative or ergative, absolutive, dative) in case-marking languages
like Russian or Dyirbal. Oblique core arguments
are marked by adpositions or oblique cases, e.g.,
instrumental and locative. Modifying each level of
Central Concepts of the Theory

Clause Structure
RRG rejects the standard formats for representing clause structure (grammatical relations, X-bar
Figure 1 Organization of role and reference grammar.
642 Role and Reference Grammar
Figure 2 The layered structure of the clause.
the clause is a periphery, which contains adjunct

modifiers, both phrasal (PPs or clauses, modifying
the core and clause) and nonphrasal (adverbs, modifying all three layers). These aspects of the LSC are
universal.
Some language have a precore slot (PrCS), which
is the position of WH-words in languages such as
English and Icelandic, and a left-detached position,
(LDP), which is the position of the preclausal element
in a left-dislocation construction. In addition, some
verb-final languages have a postcore slot (PoCS) (e.g.,
Japanese; Shimojo, 1995), and some languages also
have a right-detached position (RDP), which is the
position of the postclausal element in a right-dislocation construction. Each of the major layers (nucleus,
core, clause) is modified by one or more operators,
which include grammatical categories such as tense,
aspect, modality, and evidentiality. The LSC applies
equally to fixed word-order and free word-order languages, to head-marking and dependent-marking languages, to languages with and without grammatical
relations. It is assumed that noun phrases and adpositional phrases have a comparable layered structure;
operators in the NP include determiners, quantifiers
and adjectival and nominal modifiers. In the formal
representation of the LSC, operators are represented
in a distinct projection of the clause from the predicates and arguments (the constituent projection). This
is presented in Figures 25. In Figure 2, the peripheries have been omitted from this diagram for the
Figure 3 The LSC in English.
sake of simplicity. In Figure 3, did is labeled both

tense and IF in the operator projection, because the
position of the tense operator signals illocutionary
force in English: core-medial tense signals declarative
IF, core-initial (precore) tense signals interrogative IF,
and the absence of tense in a matrix core signals
imperative IF.
The operator projections of the clause have been
omitted in Dyirbal (Dixon, 1972) and Lakhota examples. The Dyirbal and English sentences in Figure 4
are translations of each other. The lines connecting
the determiners to the head nouns are the operator
projection within the NP, analogous to the operator
Figure 4 The LSC in Dyirbal and English.
Figure 5 The LSC in Lakhota (head-marking) and English (dependent-marking).
projection within the clause, as in Figures 23. In

head-marking languages such as Lakhota, the bound
pronominals on the verb are considered to be the
core arguments; overt NPs are within the clause in
apposition to them. Note that despite the differences
between the three languages in Figures 45, comparable structural relations, e.g., core argument and
peripheral adjunct, are represented in the same way.
It should be noted that these representations are
not abstract, unlike relational networks or functional structures; they are intended to be concrete,
in the sense that they should represent the actual
form of the sentence, including the linear sequence
of its constituent elements and their morphological
properties. The representation may be abstract
with respect to phonology or morphophonology,
e.g., the output could be in terms of abstract morphophonological units rather than concrete phonetic
ones.
Representations of constituent projections such
as these should be viewed as syntactic templates,
the inventory of which in a language constitutes an
important component of its grammar. It may be termed
the syntactic inventory and complements the lexicon.
The three central components of the LSC also

turn out to be the three fundamental building blocks
of complex sentences in human language. The unmarked pattern for the construction of complex
sentences involves combining nuclei with nuclei,
cores with cores, clauses with clauses, or sentences
with sentences. These are called levels of juncture in
RRG, i.e., nuclear juncture, core juncture, clausal
juncture, and sentential juncture. Sentential junctures are complex constructions made up of multiple
sentences, while clausal junctures involve sentences
containing multiple clauses. Examples of nuclear
junctures from French, English, and Mandarin
Chinese are given in (1) and the representation of
(1a) is in Figure 6. Justifications for these structures
can be found in Van Valin (2005).
(1a) Je
ferai
manger
1sg make.FUT eat
les ga teaux a` Jean.
the cakes to John
I will make John eat the cakes.
[two nuclei, faire and manger, in a single core]
(1b) John forced open the door.
[two nuclei, push and open, in a single core]

(1c) Ta qia o po` le y ge fa`nwan
3sg hit break PRFV one CL bowl
He broke (by hitting) a ricebowl.
[two nuclei, qiao hit and po` break, in a single
core] (Hansell, 1993)
Core junctures involve two or more cores (which

may themselves be internally complex) in a clause.
Examples from French, English, and Mandarin are

given in (2), and the structure of (2a) is presented in
Figure 7. In this type of core juncture, the two cores
share a core argument; sharing a core argument is
defined formally in terms of the linking algorithm
mapping syntactic and semantic representations into
each other.
(2a) Je
laisserai Jean manger
1sg let.FUT
John eat
I will let John eat the cakes.
les gateaux.
the cakes
(2b) I ordered Fred to force the door open.

(2c) Ta jiao wo xe z`
3sg teach 1sg write characters
She teaches me to write characters.
Figure 6 The structure of (1a).
Figure 7 The structure of (2a).
Of equal importance in the RRG theory of complex

sentences is the set of possible syntactic and semantic
relations between the units in a juncture; the
semantic relations are discussed in a related article
(see Semantics in Role and Reference Grammar). The
syntactic relations between units are called nexus
relations in RRG. Traditionally, only two basic
nexus relations are recognized, coordination and
subordination. Subordination is divided into two
subtypes, daughter subordination and peripheral
subordination. They are illustrated in Figure 8.
The embedded clause in the first sentence is a
daughter of the core node, while in the second
the embedded clause is an adjunct in the periphery
modifying the core.
In addition to distinguishing two types of subordination, RRG, following Olsons (1981) analysis of
clause linkage in Barai (a Papuan language), posits a
third nexus type: co-subordination, which is an essentially tight, dependent coordination. The dependence
is operator dependence; that is, in co-subordination,
the units obligatorily share one or more operators
at the level of juncture. In the Mandarin example in
(2c), aspect obligatorily has scope over both nuclei,
Figure 8 Daughter and peripheral subordination at the core level in English.
and therefore the nexus is co-subordination. This is

represented as in Figure 9.
The following examples from Turkish (Watters,
1993) illustrate obligatory operator sharing and the
lack of it in Turkish core co-subordination and coordination, respectively. The term coordination here is
being used for an abstract linkage relation referring to
a relationship of equivalence and operator independence at the level of juncture. It is distinct from conjunction, which is a construction type of the general
form X conj Y, which may be one of the formal
instantiations of coordinate nexus (Figure 10).
(3a) Core co-subordination
Gid-ip
go r-meli-yiz.
go-CMPL see-MODAL-1pl
We ought to go and see.
(3b) Core coordination
Mu zik dinle-yerek, uyu-yabil-ir-im.
musiclisten-CMPL sleep-MODAL-AOR-1sg
While listening to music, I can sleep.
In (3a), the modal operator -mElI- ought has

scope over both cores, and therefore the nexus is
co-subordinate; in (3b), on the other hand, the
modal operator -yAbIl- able has scope only over
the final core, hence coordinate nexus. The following
sentences from Kewa (Franklin, 1971) are a minimal
triple for the three nexus types at the clause level.
Figure 9 Nuclear co-subordination in Mandarin.
Figure 10 Turkish core junctures.
(4a) Coordination
Nipu pu-la
pare
3sg
come-3sg PRES but
n
paala
na-pa
1sg afraid NEG-be.1sg PRES
He is coming, but I am not afraid.

(4b) Co-subordination
(N) E po
la -ri
e pa-wa
(1sg) whistle say-SIM.SS come-1sgPAST
I whistled while I came, or I came whistling.
(4c) Subordination (peripheral)
(N)
E pola -to-pulu
(1sg) whistle say-1sg PRES-CAUSAL
irikai e pa-lia
dog
come-3sg FUT
Because I am whistling, the dog will come.
The four levels of juncture combine with the three

nexus types to generate eleven possible complex sentence types; there is no sentential co-subordination,
because there are no sentence-level operators, hence
no possible operator sharing. In addition, both subtypes of subordination are possible at the clause, core,
and nuclear levels. Not all of them are instantiated in
every language. English, for example, has all except
Figure 11 Interclausal syntactic relations hierarchy.
Figure 12 Predicate focus construction in English.
for nuclear subordination and coordination. The

juncture-nexus types found in a language may be
realized by more than one formal construction type;
for example, both Mary sat playing the guitar
and Robin tried to open the door instantiate core
co-subordination, while both For Sam to leave now
would be a mistake and Lisas losing her job shocked
everyone instantiate core subordination in English.
The juncture-nexus types may be ordered into a hierarchy in terms of the tightness of the syntactic link
between the units. This is given in Figure 11.
Semantic Structure
see Semantics in Role and Reference Grammar.

Focus Structure
The issue of the distribution of information in clauses

and sentences was not addressed in early work in
RRG, and Lambrechts (1994) theory of focus structure has been integrated into RRG. Focus structure is
the grammatical system that serves to indicate the
scope of the assertion in an utterance in contrast to
the pragmatic presupposition, and it is vital to the
RRG analysis of many grammatical phenomena,
e.g., constraints on pronominalization (Van Valin
and LaPolla, 1997: Sect. 5.6), the interpretation of
quantifier scope (Van Valin and LaPolla, 1997: Sect.
5.5; Van Valin, 2005: Sect. 3.6), the origin of VPs in
languages that have them (Van Valin, 2005: Sect.
3.5), and extraction restrictions (Van Valin, 1995,
1998; Van Valin and LaPolla, 1997: Sect. 9.5; Van
Valin, 2005: Sect. 7.6). An innovation in RRG is the
distinction between the potential focus domain
(PFD), i.e., the syntactic domain in the sentence
where focus may fall, and the actual focus domain,
i.e., the part that is focused in a particular sentence.
Languages vary in terms of how the PFD is restricted,
both in simple sentences and in complex sentences,
and this variation underlies important grammatical
differences across languages (Van Valin and LaPolla

1997: Sect. 5.3, 1997: Sect. 7.6). The focus structure
of an utterance is represented in a distinct projection
of the clause from the operator and constituent
projections; this is exemplified in Figure 12 for a
predicate focus construction in English. Predicate
focus is Lambrechts term for the traditional topiccomment structure with a topical subject and a focal
predicate; IU stands for basic information unit.
It is possible to represent all three projections in a
single representation, as in Figure 13.
Grammatical Relations and Linking
In the earliest work on RRG, it was argued that

grammatical relations such as subject and direct object are not universal and cannot be taken as the basis
for adequate grammatical theories. In place of these
notions, RRG employs the notion of privileged syntactic argument (PSA), which is a construction-specific relation and is defined as a restricted neutralization
of semantic roles and pragmatic functions for syntactic purposes. The other arguments in a clause are
characterized as direct or oblique core arguments;
there is nothing in RRG corresponding to direct or
indirect object.
PSAs may be characterized functionally as controllers or pivots. These two functions are exemplified in
(5) and (6).
(5a) The tall mani hit
Williamj and then
CONTROLLER
(5b) Williami was hit by
the tall manj
and then
CONTROLLER
___i/*j ran away.

PIVOT
___i/*j ran away.
PIVOT
Figure 13 Clause structure with constituent, operator, and

focus structure projections.
(6a) Bill persuaded the

tall mani
CONTROLLER
(6b) The tall mani was
persuaded by Bill
CONTROLLER
[___i to visit Sam].

PIVOT
[___i to visit Leslie].
PIVOT
Pivots are canonically the missing argument in a

construction, as in (5) and (6), while controllers prototypically supply the interpretation for a pivot. It
should be noted that there can be pivots without
controllers, e.g., the extracted element in an extraction construction, and controllers without pivots,
e.g., agreement controllers. A further contrast is
highlighted in (5) and (6), the contrast between syntactic and semantic pivots and controllers. In the
construction in (5), the controller is the first NP in
the core, the traditional subject, regardless of its semantic function, whereas in the construction in (6),
the controller is the undergoer argument, regardless
of its syntactic status. Hence the controller in (5) is
a syntactic controller, while the controller in (6)
is a semantic controller. The types of pivots and
controllers that the constructions of a language have
are typologically very significant.
The linking system relating semantic and syntactic
representations is summarized in Figure 14. Syntactic
functions such as PSA and direct core argument represent the syntactic pole of the system, while logical
structures represent the semantic pole. In every language with grammatical relations, there is an accessibility to PSA hierarchy for multiple-argument verbs;
it is the Actor-Undergoer hierarchy interpreted from
the Actor end, i.e., arg of DO > 1st arg of do0 > 1st
arg of pred0 (x, y) > 2nd arg of pred0 (x, y) > arg of
pred0 (x). In syntactically accusative constructions in
languages such as English and German, the highest
ranking argument is the default choice for PSA,
whereas in syntactically ergative constructions in languages such as Dyirbal, it is lowest ranking argument
that is the default choice. That is, in a syntactically
accusative construction the unmarked choice for
the PSA of a transitive verb is the actor, with the
undergoer being a marked choice possible only in
a passive construction. On the other hand, in a
syntactically ergative construction, the unmarked
choice for the PSA of a transitive verb is the undergoer, with the actor being a marked choice possible
only in an antipassive construction. With an intransitive verb, the hierarchy is irrelevant, as the single
macrorole functions as PSA regardless of whether it
is actor or undergoer. It should be noted that in most
languages the highest ranking argument is not the
default choice for PSA but the only choice; such languages lack voice systems. (See Van Valin and
LaPolla, 1997: Sect. 6.5)
Figure 14 RRG linking system.
An important consideration in PSA selection is

whether it is restricted to macrorole arguments or
not. In some languages, e.g., German and Dyirbal,
only actor and undergoer arguments can function as
the PSA of a construction; nonmacrorole direct core
arguments can never function in this way. In others,
e.g., Icelandic and Georgian, nonmacrorole direct
core arguments can be selected to serve as the PSA
of a construction. This contrast is seen most clearly in
the status of dative NPs: in German they can never be
true subjects, while in Icelandic they can.
Logical structures, macroroles and the hierarchy
linking them are discussed in Semantics in Role and
Reference Grammar. This part of the system is universal, in that there is very little cross-linguistic variation; this is the domain of lexical processes, as
mentioned in that chapter. Where languages differ
substantially is how macroroles and other arguments
link into the syntax.
The reason the arrows in Figure 14 are doubleheaded is that the linking system works both from
semantics to syntax and from syntax to semantics.
A theory that could describe the linking from semantics to syntax only could be part of a language production system, but it would not be adequate for a
comprehension system. In such a system, the parser,
as an idealization, would take the input and produce
a structured syntactic representation of it, identifying
the elements of the layered structure of the clause
and the cases, adpositions, and other grammatically relevant elements in the sentence. It is then the
grammars job to map this structure into a semantic
representation, as the first step in interpreting it, and
this is where the syntax ! semantics linking algorithm is required. The details of the linking
algorithms are given in Van Valin (2005).
Most of what counts as syntax in many theories,
e.g., case assignment, agreement, and WHmovement, is handled in RRG in terms of the syntactic phase of the linking. The analysis of reflexivization
in RRG follows the approach in Jackendoff (1992)
and states the constraints for core-internal (clausebound in other theories) reflexivization at the LS
level, not with respect to the syntactic representation.
RRG treats constructions as an important part of
syntax, and they are represented in terms of constructional schemas. Cross-constructional and crosslinguistic generalizations are captured in terms of
the general principles and constraints that constitute
the linking algorithms, e.g., the ActorUndergoer
Hierarchy, the layered structure of the clause, and
the PSA selection hierarchy. Only the idiosyncratic,
language-specific features of constructions are represented in constructional schemas, which include
syntactic, morphological, semantic and pragmatic
(focus structure) information.
The linking in a WH-question in English is illustrated in Figure 15; the subscripts ACV and ACS
stand for activated and accessible, and they refer
to different cognitive statuses that a referent of the
element may have; cf. Lambrecht (1994).
Constraints on WH-question formation and other
extraction constructions are explained in terms of the
interaction of focus structure and syntax, in particular in terms of restrictions on the potential focus
domain (Van Valin, 1995, 1998, 2005).
Figure 15 Linking syntax and semantics in a simple sentence in English.
Some Implications of RRG

RRG illustrates one possible answer to the questions
stated at the beginning, and it shows that it is possible
to have a rigorous, typologically sensitive grammatical theory that takes semantics and pragmatics as
central features.
With respect to cognitive issues, RRG adopts the
criterion of psychological adequacy formulated in
Dik (1991), which states that a theory should be
compatible with the results of psycholinguistic research on the acquisition, processing, production,
interpretation and memorization of linguistic expressions (Dik, 1991: 248). Van Valin (2004) relates
RRG to psycholinguistic models of production and
comprehension, and Bornkessel et al. (2004) provide
an explanation for certain experimental results regarding German language processing using RRG.
Guest (2003) presents a parser based on RRG that
can parse a range of syntactic structures in both
English and Dyirbal. Kailuweit et al. (2003) and
Butler (2004) propose implementations of the RRG
semantics-to-syntax linking algorithm for French and
English, respectively.
The RRG approach to language acquisition,
sketched in Van Valin (1991, 1994) and Van Valin
and LaPolla (1997), rejects the position that grammar
is radically arbitrary and hence unlearnable, and
maintains that it is relatively motivated (in Saussures
sense) semantically and pragmatically. Accordingly,
there is sufficient information available to the child
in the speech to which it is exposed to enable it to
construct a grammar. For example, Braine (1992)
shows how a conception of clause structure very
much like the layered structure of the clause could
be constructed developmentally by the child. Rispoli

(1991a, 1991b, 1994) shows how the lexical representations in RRG and the conception of grammatical
relations could be learned. Bowerman (1990) provides evidence in favor of the view that rules linking
syntactic and semantic representations of the type
summarized in Figure 15 are learned, and Weist
(1990) shows how the RRG theory of grammatical
relations provides an explanation for the pattern of
development of grammatical relations in children
learning Polish. Weist et al. (2004) shows the importance of Aktionsart for the understanding of the acquisition of verbs. Van Valin (1994, 1998) puts
forward an account of how some of the constraints
on linking between syntactic and semantic representations in complex sentences (i.e., subjacency) could
be learned. Certain otherwise problematic facts about
the acquisition of WH-questions are explained in
terms of RRG in Van Valin (1998, 2002). Van Valin
(2001) explores the predictions made by the theory
of clause linkage and presents evidence from seven
languages in support of them.
See also: Clause Structure in Spoken Discourse; Grammatical Relations and Arc-Pair Grammar; Information
Structure in Spoken Discourse; Semantics in Role and
Reference Grammar; Subordination.
Bibliography
Bornkessel I, Schlesewsky M & Van Valin R D, Jr. (2004).
Syntactic templates and linking mechanisms: a new
approach to grammatical function asymmetries. Poster
presented at 17th CUNY Conference on Human Sentence
Processing. [available on the RRG web site].

Bowerman M (1990). Mapping semantic roles onto syntactic functions: are children helped by innate linking
rules? Linguistics 28, 12531289.
Braine M D S (1992). What sort of innate structure
is needed to bootstrap into syntax? Cognition 45,
77100.
Butler C (2004). Notes towards an incremental implementation of the Role and Reference Grammar semantics-to-syntax mapping rules for English. In Hannay M
& Steen G (eds.) The English clause: usage and structure.
Dik S (1991). Functional grammar. In Droste F & Joseph J
(eds.) Linguistic theory and grammatical description.
Amsterdam: John Benjamins. 247274.
Dixon R M W (1972). The Dyirbal language of north
Queensland. Cambridge: Cambridge University Press.
Franklin K (1971). A grammar of Kewa, New Guinea.
(Pacific Linguistics C-16). Canberra: Australian National
University.
Guest E (2003). Parsing for Role and Reference Grammar.
Paper presented at the 2003 International Conference on
Role and Reference Grammar, UNESP, Sa o Jose do Rio
Preto, Brazil.
Hansell M (1993). Serial verbs and complement constructions in Mandarin: a clause linkage analysis. In Van
Valin R D (ed.) Advances in Role and Reference Grammar. Amsterdam: John Benjamin. 197233.
Jackendoff R (1992). Mme. Tussaud meets the binding
theory. Natural Language and Linguistic Theory 10,
131.
Kailuweit R, Bender T, Hartung M et al. (2003). Parsing
English and French using the RRG paradigm. Paper
presented at the 2003 International Conference on Role
and Reference Grammar, UNESP, Sa o Jose do Rio Preto,
Brazil.
Lambrecht K (1994). Information structure and sentence
form. Cambridge: Cambridge University Press.
Olson M (1981). Barai clause junctures: toward a functional theory of interclausal relations. Ph.D dissertation,
Australian National University.
Rispoli M (1991a). The acquisition of verb subcategorization in a functionalist framework. First Language 11,
4163.
Rispoli M (1991b). The mosaic acquisition of grammatical
relations. Journal of Child Language 18, 517552.
Rispoli M (1994). Structural dependency and the acquisition of grammatical relations. In Levy Y (ed.) Other
children, other languages: issues in the theory of language
acquisition. Hillsdale, NJ: LEA. 265301.
Shimojo M (1995). Focus structure and morphosyntax
In Japanese: Wa and ga, and word order flexibility.
Ph.D. dissertation, SUNY at Buffalo [avaliable on RRG
web site].
Van Valin R D, Jr. (1991). Functionalist linguistic theory

and language acquisition. First Language 11, 740.
Van Valin R D, Jr. (1993a). A synopsis of role and reference grammar. In Van Valin R D (ed.) Advances in Role
and Reference Grammar. Amsterdam: John Benjamin.
1164.
Van Valin R D, Jr. (ed.) (1993b). Advances in Role and
Reference Grammar. Amsterdam: John Benjamin.
Van Valin R D, Jr. (1994). Extraction restrictions, competing theories and the argument from the poverty of the
stimulus. In Lima S et al. (eds.) The reality of linguistic
rules. Amsterdam: John Benjamin. 243259.
Van Valin R D, Jr. (1995). Toward a functionalist account
of so-called extraction constraints. In Divriendt B et
al. (eds.) Complex structures: a functionalist perspective.
Berlin: Mouton de Gruyter. 2960.
Van Valin R D, Jr. (1998). The acquisition of WH-questions
and the mechanisms of language acquisition. In
Tomasello M (ed.) The new psychology of language:
cognitive and functional approaches to language structure. Mahwah, NJ: LEA. 221249.
Van Valin R D, Jr. (2001). The acquisition of complex
sentences: a case study in the role of theory in the study
of language development. CLS 36(2), 511531.
Van Valin R D, Jr. (2002). The development of subjectauxiliary inversion in English wh-questions: an alternative analysis. Journal of Child Language 29, 161175.
Van Valin R D, Jr. (2004). On the relationship between
syntactic theory and models of language processing. In
Bornkessel I et al. (eds.) Semantic role universals and
argument linking: Theoretical typological and psycho/
neurolinguistic perspectives. Berlin: Mouton de Gruyter.
(in press).
Van Valin R D, Jr. (2005). Exploring the syntax-semantics
interface. Cambridge: Cambridge University Press.
Van Valin R D, Jr. & LaPolla R J (1997). Syntax: structure,
meaning & function. Cambridge: Cambridge University
Press.
Watters J (1993). An investigation of Turkish clause linkage. In Van Valin (ed.) Advances in Role and Reference
Grammar. Amsterdam: John Benjamin. 535560.
Weist R (1990). Neutralization and the concept of subject
in child Polish. Linguistics 28, 13311349.
Weist R, Pawlak A & Carapella J (2004). Syntacticsemantic interface in the acquisition of verb morphology.
Journal of Child Language 31, 3160.
Relevant Website
http://linguistics.buffalo.edu/research/rrg.html the RRG
web site, which contains dissertations, theses and papers
in pdf format for downloading.
Roman Ars Grammatica 651
Romaine, Suzanne (b. 1951)

M Pasquale, Cornerstone University,
Grand Rapids, MI, USA
Suzanne Romaine is a native of Massachusetts,

USA and graduated from Bryn Mawr College in
Pennsylvania with an A.B. in German and Linguistics.
She traveled to Scotland on a Rotary International
Foundation Fellowship and earned an M.Litt. in
Phonetics and Linguistics from the University of
Edinburgh, Scotland. She went on to earn her Ph.D.
at the University of Birmingham, England in 1981.
She has received honorary doctorates from the
University of Troms, Norway in 1998 and the University of Uppsala, Sweden in 1999. She received the
Canada Commonwealth Scholarship at the Ontario
Institute for Studies in Education, University of
Toronto. Romaine was named the Kristin Hesselgren
Professor at the University of Uppsala for 1991
1992. This award is granted to an outstanding
woman in the area of Humanities and is sponsored
by the Swedish Research Council for the Humanities
and Social Sciences. Since 1984, she has held the
position of Merton Professor of English at the University of Oxford.
Romaines research and publications span several
areas such as sociolinguistics, language acquisition,
bilingualism, and language loss. She has written an
introductory textbook in the area of sociolinguistics,
Language in society (1994), now in its second edition.
Her chapter on sexism and language was expanded
into the monograph Communicating gender (1999).
Romaine has conducted sociolinguistic research in
the United Kingdom and in the Pacific region. She did
fieldwork in Scotland on the language of working-class
children. She has also done research in the area of
bilingualism and language loss among the Punjabi
speakers in England. Her textbook Bilingualism
(1989; 2nd edition, 1995) relates much of her findings

in these areas along with her work with language groups in Papua New Guinea. She has also written the book Pidgin and Creole languages (1988).
Recently she has done research on Hawaiian Creole
English. A sample of her work in this area can be found
in a book she edited with John Rickford entitled Creole
genesis, attitudes and discourse (1999).
Romaine has earned wide acclaim for her book on
the state of language loss called Vanishing voices,
written with Daniel Nettle (2000). The book has
been translated into several languages, including
Italian, French, Japanese, and Turkish. The volume
earned the British Association of Applied Linguistics
book-of-the-year award in 2001.
See also: Bilingualism; Language Acquisition Research
Methods; Pidgins and Creoles: Overview.
Bibliography
Nettle D & Romaine S (2000). Vanishing voices: The
extinction of the worlds languages. Oxford: Oxford
University Press.
Rickford J R & Romaine S (1999). Creole genesis, attitudes
and discourse: studies celebrating Charlene J. Sato.
Romaine S (1984). The language of children and adolescents: the acquisition of communicative competence.
Oxford: Blackwell.
Romaine S (1988). Pidgin and creole languages. London:
Longman.
Romaine S (1989). Bilingualism (2nd edn., 1995). Oxford:
Blackwell.
Romaine S (1994). Language in society: An introduction to
sociolinguistics (2nd edn.). Oxford: Oxford University
Press.
Romaine S (1999). Communicating gender. Mahwah, NJ:
Lawrence Erlbaum.
Roman Ars Grammatica

E Hovdhaugen, University of Oslo, Oslo, Norway
The Roman ars grammatica comprises Roman grammatical works written between the 3rd and the 6th
century A.D. They are all intended as textbooks in the
schools or as handbooks for teachers. The majority
consists of descriptions of Latin morphology focusing

on the word classes (parts of speech), but many of
the works also include treatments of phonology,
metrics, and stylistics. The standard textbooks of
the Roman ars grammatica (especially Donatuss
two grammars from the 4th century A.D.) were to
form the basis for grammar teaching and grammar
writing in Europe for more than a thousand years.
Romaine, Suzanne (b. 1951)

M Pasquale, Cornerstone University,
Grand Rapids, MI, USA
Suzanne Romaine is a native of Massachusetts,

USA and graduated from Bryn Mawr College in
Pennsylvania with an A.B. in German and Linguistics.
She traveled to Scotland on a Rotary International
Foundation Fellowship and earned an M.Litt. in
Phonetics and Linguistics from the University of
Edinburgh, Scotland. She went on to earn her Ph.D.
at the University of Birmingham, England in 1981.
She has received honorary doctorates from the
University of Troms, Norway in 1998 and the University of Uppsala, Sweden in 1999. She received the
Canada Commonwealth Scholarship at the Ontario
Institute for Studies in Education, University of
Toronto. Romaine was named the Kristin Hesselgren
Professor at the University of Uppsala for 1991
1992. This award is granted to an outstanding
woman in the area of Humanities and is sponsored
by the Swedish Research Council for the Humanities
and Social Sciences. Since 1984, she has held the
position of Merton Professor of English at the University of Oxford.
Romaines research and publications span several
areas such as sociolinguistics, language acquisition,
bilingualism, and language loss. She has written an
introductory textbook in the area of sociolinguistics,
Language in society (1994), now in its second edition.
Her chapter on sexism and language was expanded
into the monograph Communicating gender (1999).
Romaine has conducted sociolinguistic research in
the United Kingdom and in the Pacific region. She did
fieldwork in Scotland on the language of working-class
children. She has also done research in the area of
bilingualism and language loss among the Punjabi
speakers in England. Her textbook Bilingualism
(1989; 2nd edition, 1995) relates much of her findings

in these areas along with her work with language groups in Papua New Guinea. She has also written the book Pidgin and Creole languages (1988).
Recently she has done research on Hawaiian Creole
English. A sample of her work in this area can be found
in a book she edited with John Rickford entitled Creole
genesis, attitudes and discourse (1999).
Romaine has earned wide acclaim for her book on
the state of language loss called Vanishing voices,
written with Daniel Nettle (2000). The book has
been translated into several languages, including
Italian, French, Japanese, and Turkish. The volume
earned the British Association of Applied Linguistics
book-of-the-year award in 2001.
See also: Bilingualism; Language Acquisition Research
Methods; Pidgins and Creoles: Overview.
Bibliography
Nettle D & Romaine S (2000). Vanishing voices: The
extinction of the worlds languages. Oxford: Oxford
University Press.
Rickford J R & Romaine S (1999). Creole genesis, attitudes
and discourse: studies celebrating Charlene J. Sato.
Romaine S (1984). The language of children and adolescents: the acquisition of communicative competence.
Oxford: Blackwell.
Romaine S (1988). Pidgin and creole languages. London:
Longman.
Romaine S (1989). Bilingualism (2nd edn., 1995). Oxford:
Blackwell.
Romaine S (1994). Language in society: An introduction to
sociolinguistics (2nd edn.). Oxford: Oxford University
Press.
Romaine S (1999). Communicating gender. Mahwah, NJ:
Lawrence Erlbaum.
Roman Ars Grammatica

E Hovdhaugen, University of Oslo, Oslo, Norway
The Roman ars grammatica comprises Roman grammatical works written between the 3rd and the 6th
century A.D. They are all intended as textbooks in the
schools or as handbooks for teachers. The majority
consists of descriptions of Latin morphology focusing

on the word classes (parts of speech), but many of
the works also include treatments of phonology,
metrics, and stylistics. The standard textbooks of
the Roman ars grammatica (especially Donatuss
two grammars from the 4th century A.D.) were to
form the basis for grammar teaching and grammar
writing in Europe for more than a thousand years.
652 Roman Ars Grammatica
Background
The term ars grammatica means the art of grammar
and is a translation of Greek techne grammatike with
the same meaning. It is also likely that the Roman ars
grammatica originated from the Greek genre (see
Dionysius Thrax and Hellenistic Language Scholarship), but owing to the lack of surviving Greek grammatical texts and the dating problems within Greek
linguistics, the exact nature of the connection between the two grammatical traditions remains unclear.
Just as in the Greek world, the Roman grammatical
textbooks are designed for providing a metalanguage
and a method for analyzing those literary texts (mainly metrical texts like the works of Homer and Virgil),
which were used as the principal textbooks in the
schools. Gradually, however, the textbooks became
more linguistic, and literary considerations faded into
the background. There are two likely reasons for this
development:
1. The increasing divergence between the spoken and
written language created a need for teaching a
standardized literary language in the schools.
2. The establishment of an organized school system
combined with a professionalism among the teachers, who preferred a fixed curriculum.
The works labeled ars grammatica cover a wide
variety of subjects and approaches to language and
are, in comparison with the few reminiscences of
Greek school grammars from antiquity, impressive
both in quantity and quality. The extant grammatical
texts, however, represent but a part of the Roman
grammatical literature. We have numerous references in Roman literature to grammarians otherwise
unknown to us or to grammars now lost. Furthermore, the names of the authors of many of the extant
works are unknown or uncertain (cf. Kaster, 1988). In
some cases a grammar is falsely attributed to an otherwise famous person known to have written an ars
grammatica (e.g., Remmius Palaemon) or to a wellknown person (e.g., St Augustine). It was common for
the grammarians to copy each other extensively, often
without acknowledging their sources.
The Roman Grammarians

The first extant ars grammatica is the work in three
books by Marius Plotius Sacerdos (3rd century A.D.).
The work is badly organized and apparently starts
with prepositions and verbs (the manuscript is fragmentary in the beginning). After the parts of speech
have been discussed, there is an extensive treatment
of the ablative case and a possible seventh case in
Latin. The second book is devoted to syllable quantity

and phonotactic matters as well as to a classification
of stems and words according to their final sounds
and syllables. The third book is entirely devoted to
metrics.
The African grammarian Flavius Sosipater Charisius
(4th century A.D.) wrote an extensive ars grammatica in
five books, which treats morphology (including an
extensive classification of words into inflectional classes), stylistics, and metrics and which also includes a
short description of the letters and their pronunciation, a list of synonyms, a list of words with different
genders in Greek and Latin, and so forth. A characteristic feature of Charisiuss work, and one which is
unique in the ars grammatica tradition, is that on most
controversial points he quotes extensively the views of
other authors (many of them otherwise unknown). To
some extent, his grammar is a compilation of early
Roman grammatical literature.
Diomedes (4th or 5th century A.D.) wrote an ars
grammatica in three volumes, which in many respects
follows the arrangement of Sacerdos, except that the
second book contains just a short sketch of phonology and then treats stylistics and poetical matters,
including cases of correct and incorrect Latin.
The most influential Roman grammarian of the 4th
century was Aelius Donatus, who wrote two grammars: Ars minor and Ars major. The Ars minor only
treats the word classes and is written as a dialogue
between the teacher who asks and the pupil who
answers. The Ars major also includes a short phonology and aspects of correct and incorrect Latin. Ars
major contains 21 chapters (of these topics only 715
are covered by the Ars minor).
The ars grammatica attributed to Marius Victorinus (4th century A.D.) illustrates another type of ars
grammatica. Victorinus discusses in his preface what
grammar should be and stresses the literary aspect of
his grammar and its importance for literary studies,
but limits his treatment to phonology, orthography,
and metrical aspects of the syllable. A similar
approach is found in the ars attributed to Atilius
Fortunatianus (4th century A.D.?), which is mainly a
book on metrics, but which in the beginning contains
some short chapters on vowels, consonants, and syllables. Victorinus also wrote a short elementary ars in
the form of questions and answers (a format that
apparently started with Dontuss Ars minor and
which became very popular in late antiquity and the
Middle Ages), but which unlike Donatus Ars minor,
contains a lot of Greek examples.
Consentius (date uncertain) and Phocas (5th century
A.D.) both limited their artes to the noun and the
verb. Consentius also wrote a separate treatise about
grammatical and stylistic errors.
The teaching of grammar beyond the elementary

stage usually consisted in the teachers reading a paragraph from a grammar and then commenting on it.
Some teachers published their commentaries, thereby
creating a new grammatical genre demonstrating the
versatility and originality of Roman linguistics. Most
influential were the commentaries on Donatuss Ars
major by Marius Servius Honoratus (4th century A.D.)
and by Pompeius (5th century A.D.). Pompeius represents a semantically oriented approach to grammatical analysis, which already can be found implicit in
Servius (on whom Pompeius, to a large extent, based
his work) and Phocas, and which became central in
Priscian and medieval linguistics.
In contrast to the monolingual Greek grammar,
Latin grammar was bilingual, covering both Latin
and Greek. Many authors of an ars grammatica neglected Greek, however, and had no Greek examples
in their texts (e.g., Donatus, Sacerdos) while others
(e.g., Charisius, Diomedes, and Priscian) extensively
compared their analysis of Latin with Greek and quoted Greek examples. Two authors also wrote contrastive studies. Ambrosius Theodosius Macrobius
(ca. 400 A.D.) analyzed differences and similarities between the verb in Latin and Greek, focusing on verb
morphology, while Dositheus (4th century A.D.) wrote a
bilingual ars grammatica.
Priscian (see Priscianus Caesariensis (d. ca. 530))
represents both the culmination and the end of
Roman linguistics. He became one of the most influential grammarians in history, and his impressive
work, the Institutiones grammaticae in 18 books,
goes far beyond what other Roman grammarians
had aimed at or achieved.
The Roman Ars Grammatica An Attempt

at a Synopsis
The Roman grammatical tradition was a very versatile and multifaceted one characterized by continuity
and development. On the one hand, there is the formally oriented elementary school grammar tradition
exemplified by Donatuss two grammars and by
numerous similar grammars, mostly by unknown
authors. This tradition is undoubtedly very uniform,
and the differences between the various grammars are
relatively small. With the collapse of the Roman
Empire there was a general decline in higher education and centers of learning in Western Europe.
Donatuss elementary textbooks or Christianized
versions of them became the basis for elementary
teaching of Latin in Western Europe (Holtz, 1981).
In a situation where the language of the pupils was
either very divergent from the norms of classical Latin
or another language altogether, these books may
also have been the most suitable ones available. But

in the classical Roman educational system, where
the language of the pupils was quite close to the
written language and where the main purpose of
grammar teaching was to help the pupils to acquire
a necessary metalanguage for reading and analyzing
poetical literary texts, grammars like those of Victorinus and Fortunatianus must have been much more
instrumental.
Other grammars (e.g., those of Sacerdos, Charisius,
Diomedes, and Priscian) and the commentaries on
Donatus cover a number of different genres, topics,
and theoretical approaches. Usually they include, besides the basic morphological part, a treatment of metrical matters, phonology (especially syllable quantity),
orthography, and a comparison of Latin and Greek.
A characteristic aspect of the Roman grammatical
tradition is the extensive treatment of stylistic matters,
including statements concerning correct and incorrect Latin. Much space was devoted to establishing
principles for correctness of language and to asking
whether structural uniformity of paradigms, language
usage, and/or the authority of the best authors should
be the decisive criterion in cases of doubt.
While in the beginning the grammarians used a
mixture of semantic and formal morphological criteria in their analysis and classification of Latin morphology, Donatus established a purely nonsemantic
approach, which became dominant in the elementary
ars grammatica. Among other grammarians, however, a more semantically oriented approach emerged,
culminating in Pompeius, Phocas, and Priscian, who
laid the foundation for the philosophical approach to
language analysis in the Middle Ages.
Most authors had their favorite theme(s); for instance, the Instituta artium spuriously attributed to
Probus concentrates on the comparison of adjectives.
Another popular topic was the analysis of the Latin
case system. Not only had Latin one more case than
Greek (namely, the ablative), but also the semantic
and partly morphological heterogeneity of the ablative (originating from three Indo-European cases, viz.
instrumental, locative, and ablative) led some grammarians (e.g., Sacerdos and Diomedes) to discuss
the position of the ablative and to ask whether the
ablative was, in effect, one or two cases.
While the phonology (i.e., letters and their pronunciation and syllable quantity) and the morphology
with its numerous paradigms and lists of examples
are based in most instances on data without sources
and probably to a large extent on the authors introspection (Priscian is an exception), stylistic and metrical matters are usually well-documented with
quotations from the main Latin poets (very rarely
prose writers), especially Virgil and Horace.
The most important aspect of the Roman ars grammatica is that it formed the grammatical metalanguage and, above all, the structural arrangement for
elementary and advanced grammars and textbooks
for language teaching in Western Europe for more
than 1000 years. Even today many grammars have
clear echoes of Donatus.
See also: Dionysius Thrax and Hellenistic Language
Scholarship; Priscianus Caesariensis (d. ca. 530).
Bibliography
Barwick K (1922). Remmius Palaemon und die ro mische
ars grammatica. Leipzig: Dietrich.
Holtz L (1981). Donat et la tradition de lenseignement

grammatcal: E tude sur lArs Donati et sa diffusion (IVeIXe sie`cle) et e dition critique. Paris: Centre National de la
Reherche Scientifique.
Hovdhaugen E (1986). Genera verborum quot sunt: Observations on the Roman grammatical tradition. Historiographia linguistica XIII(2/3), 307321.
Kaster R A (1988). Guardians of language: the grammarian
and society in Late Antiquity. Berkeley, CA: University of
California Press.
Keil H (ed.) (185580). Grammatici Latini (7 vols & supplement). Leipzig: Teubner.
Rosier I (ed.) (1988). Lhe ritage des grammairiens latina de
lantiquite aux lumie`res. Louvain: Peeters.
Schmitter P (ed.) (1991). Sprachteorien der abendla ndischen
Antike. Tu bingen: Narr.
Romance Languages
R Wright, University of Liverpool, Liverpool, UK
The Romance languages are those that have developed from the spoken Latin of the early Middle Ages.
In this sense one can claim that Latin is not dead;
about a quarter of the worlds population still speak
it; but it has acquired several new geographically
based names (Spanish, Portuguese, French, Italian,
Romanian, Catalan, Occitan, Sardinian, Galician,
Rhaeto-romanic, here listed roughly in descending
order of number of speakers). These are for political
reasons considered to be separate Romance languages, but there is still essentially one dialect continuum overlaid by the several artificially extended
standards. Apart perhaps from Romanian, the location and history of whose earliest speakers is still
controversial, the definitive divergence into separately identifiable languages should be dated to no earlier
than the 9th century, and in several cases later.
Reconstruction
There are two main kinds of evidence for the
Romance (spoken Latin) that existed before the separate languages diverged: surviving written texts and
the results of reconstruction. Hall and others have
reconstructed a hypothetical Proto-Romance on the
basis of the later Romance languages; features they
have in common are taken to have existed in their
ancestors. As compared with classical Latin, this
Proto-Romance contains, for example, no neuter
nouns, no ablative cases, no datives and genitives
outside pronouns, no synthetic passives or futures,

no adverbs in -iter, no phonemic length distinctions
in vowels, no originally final consonants other
than alveolars, and no velar consonants before front
vowels other than those that were originally labiovelar. On the other hand, the evidence of modern
Romance languages suggests that their base included
extended uses of prepositions (particularly ad and de
to replace inflectional nominal suffixes); analytic
passives with auxiliary esse tense-indeterminate
participles; extended use of grammatically reflexive
se with passive meaning; analytic futures (and
conditionals) formed with the infinitive habeo;
new analytic perfects (including future perfects
and pluperfects) formed with activized participles habeo; extensive use of ille and ipse with the
functions of the definite article; many diminutives
in -iculum and other affixed forms (such as adiutare,
rather than iuvare, as the base of Port ajudar, Sp
ayudar, Cat ajudar, Fre aider, Italian aiutare,
Romanian ajuta, etc.); the use of preposed magis or
plus instead of comparative -ior; palatal affricates
and semivowels; and much new vocabulary from, in
particular, Germanic sources.
Texts
This reconstructed language is not very much like that

of the surviving written texts of the time. Janson
described reconstruction and textual analysis as two
different key-holes through which one can look into
the same large room, for the rules of correct writing
did not change, and mistakes are rarely attested.
Most texts were written on perishable wax tablets
or papyruses; the extant versions are usually later
The most important aspect of the Roman ars grammatica is that it formed the grammatical metalanguage and, above all, the structural arrangement for
elementary and advanced grammars and textbooks
for language teaching in Western Europe for more
than 1000 years. Even today many grammars have
clear echoes of Donatus.
See also: Dionysius Thrax and Hellenistic Language
Scholarship; Priscianus Caesariensis (d. ca. 530).
Bibliography
Barwick K (1922). Remmius Palaemon und die romische
ars grammatica. Leipzig: Dietrich.
Holtz L (1981). Donat et la tradition de lenseignement

grammatcal: Etude sur lArs Donati et sa diffusion (IVeIXe sie`cle) et edition critique. Paris: Centre National de la
Reherche Scientifique.
Hovdhaugen E (1986). Genera verborum quot sunt: Observations on the Roman grammatical tradition. Historiographia linguistica XIII(2/3), 307321.
Kaster R A (1988). Guardians of language: the grammarian
and society in Late Antiquity. Berkeley, CA: University of
California Press.
Keil H (ed.) (185580). Grammatici Latini (7 vols & supplement). Leipzig: Teubner.
Rosier I (ed.) (1988). Lheritage des grammairiens latina de
lantiquite aux lumie`res. Louvain: Peeters.
Schmitter P (ed.) (1991). Sprachteorien der abendlandischen
Antike. Tubingen: Narr.
Romance Languages
R Wright, University of Liverpool, Liverpool, UK
The Romance languages are those that have developed from the spoken Latin of the early Middle Ages.
In this sense one can claim that Latin is not dead;
about a quarter of the worlds population still speak
it; but it has acquired several new geographically
based names (Spanish, Portuguese, French, Italian,
Romanian, Catalan, Occitan, Sardinian, Galician,
Rhaeto-romanic, here listed roughly in descending
order of number of speakers). These are for political
reasons considered to be separate Romance languages, but there is still essentially one dialect continuum overlaid by the several artificially extended
standards. Apart perhaps from Romanian, the location and history of whose earliest speakers is still
controversial, the definitive divergence into separately identifiable languages should be dated to no earlier
than the 9th century, and in several cases later.
Reconstruction
There are two main kinds of evidence for the
Romance (spoken Latin) that existed before the separate languages diverged: surviving written texts and
the results of reconstruction. Hall and others have
reconstructed a hypothetical Proto-Romance on the
basis of the later Romance languages; features they
have in common are taken to have existed in their
ancestors. As compared with classical Latin, this
Proto-Romance contains, for example, no neuter
nouns, no ablative cases, no datives and genitives
outside pronouns, no synthetic passives or futures,

no adverbs in -iter, no phonemic length distinctions
in vowels, no originally final consonants other
than alveolars, and no velar consonants before front
vowels other than those that were originally labiovelar. On the other hand, the evidence of modern
Romance languages suggests that their base included
extended uses of prepositions (particularly ad and de
to replace inflectional nominal suffixes); analytic
passives with auxiliary esse tense-indeterminate
participles; extended use of grammatically reflexive
se with passive meaning; analytic futures (and
conditionals) formed with the infinitive habeo;
new analytic perfects (including future perfects
and pluperfects) formed with activized participles habeo; extensive use of ille and ipse with the
functions of the definite article; many diminutives
in -iculum and other affixed forms (such as adiutare,
rather than iuvare, as the base of Port ajudar, Sp
ayudar, Cat ajudar, Fre aider, Italian aiutare,
Romanian ajuta, etc.); the use of preposed magis or
plus instead of comparative -ior; palatal affricates
and semivowels; and much new vocabulary from, in
particular, Germanic sources.
Texts
This reconstructed language is not very much like that

of the surviving written texts of the time. Janson
described reconstruction and textual analysis as two
different key-holes through which one can look into
the same large room, for the rules of correct writing
did not change, and mistakes are rarely attested.
Most texts were written on perishable wax tablets
or papyruses; the extant versions are usually later
Romance Languages 655
manuscript copies prepared by scribes who had specific instructions to correct their originals according
to the arcane and eventually archaic rigidities of the
Imperial grammarians. Texts without such distortions are few; Adams has published some letters and
drafts, and Vaananens study of the Pompeii Graffiti
revolutionized the discipline by showing how incorrectly nonscholars wrote in 79 A.D. Even these texts,
however, are obviously not phonetic transcriptions of
actual speech. From painstaking statistical analyses of
surviving inscriptions (mostly on tombstones), whose
textual details cannot be corrected, Herman has
concluded that Imperial spoken Latin was evolving
but also converging, with new features starting in one
place becoming eventually attested anywhere. Some
further progress is made by studying borrowings from
spoken Latin into, for example, Irish, Welsh, Berber,
Albanian, and Greek.
Divergence
Wide variation arose, but this need not imply mutual
unintelligibility. Many historians, textual critics, philologists, sociolinguists, and historical linguists currently view early Medieval Romance Europe as a
single lively speech community, where almost everyone could understand old-fashioned written texts
when read aloud (McKitterick, 1990, Wright, 1991).
These were not Dark Ages. Early Medieval speakers
rarely made metalinguistic distinctions that we take
for granted now, neither diatopic (between French,
Spanish, etc.) nor diastratic (between Romance and
Medieval Latin). The latter distinction was probably
imported from Germanic-speaking areas, where vernacular Germanic and official Latin grammatica
were unrelated and self-evidently different languages;
conscious distinctions between separate Romance
languages became widespread only after the fashion
for inventing distinctive writing systems in different
areas, which began experimentally in 9th-century
eastern France but generalized only in the 12th and
13th centuries. Indeed, to some extent the speech
of the central Romance area is still mutually intelligible, given goodwill and clarity from those in the
conversation; peripheral languages, such as Romanian, Portuguese, and French, are rarely intelligible
elsewhere.
See also: Catalan; French; Italian; Occitan; Portuguese;
Rhaeto Romance; Romanian; Spanish.
Bibliography
Adams J N (1977). The vulgar Latin of the letters of Claudius
Terentianus. Manchester: Manchester University Press.
Adams J N (2003). Bilingualism and the Latin language.
Banniard M (1992). Viva Voce: communication e crite et
communication orale du IVe au IXe sie`cle en Occident
latin. Paris: Institut des Etudes Augustiniennes.
Banniard M (ed.) (2002). Langages et peuples dEurope.
Toulouse-le-Mirail: CNRS.
Hall R A (1976). Proto-Romance phonology. New York:
Elsevier.
Harris M & Vincent N (eds.) (1987). The Romance languages. London: Croom Helm.
Herman J (1990). Du latin aux langues romanes. E tudes de
linguistique historique. Tubingen: Niemeyer.
Herman J (ed.) (1998). La transizione dal latino alle lingue
romanze. Tubingen: Niemeyer.
Herman J (2000). Vulgar Latin. Pennsylvania: Penn State
Press.
Janson T (1979). Mechanisms of language change in Latin.
Stockholm: Almqvist and Wiksell.
McKitterick R (ed.) (1990). The uses of literacy in early
Medieval Europe. Cambridge: Cambridge University
Press.
Selig M (ed.) (1993). Le passage a` lecrit des langues
romanes. Tubingen: Narr.
Trotter D A & Gregory S (eds.) (1997). De mot en mot:
aspects of medieval linguistics. Cardiff: University of
Wales Press.
Vaananen V (1937). Le latin vulgaire des inscriptions pompe iennes. Helsinki: Academy.
Wright R (ed.) (1991). Latin and the Romance languages in
the early Middle Ages. London: Routledge.
Wright R (1995). Early Ibero-Romance. Newark, DE: Juan
de la Cuesta.
Wright R (2003). A sociophilological study of late Latin.
Turnhout: Brepols.
656 Romani
Romani
Y Matras, University of Manchester, Manchester, UK
Definitions
Romani (referred to by its speakers as romani chib
the Romani language or romanes in a Romani way)
is the only Indo-Aryan language spoken exclusively
in Europe, as well as by emigrant populations in
the Americas and Australia. The language is often
referred to as Gypsy; it is important, however, to
distinguish between Romani, which is the fully
fledged, everyday family and community language
spoken by the people who call themselves Rom, and
secret or in-group vocabularies employed in various
parts of the world, including in Europe, by other
populations of peripatetics or so-called servicenomads. There is nevertheless some interface between
the two phenomena: in some regions of Europe, especially the western margins (Britain, the Iberian peninsula, Scandinavia), Romani-speaking communities
have given up their language in favor of the majority
language but have retained Romani-derived vocabulary as an in-group code. Such codes, for instance
Angloromani (Britain), Calo (Spain), or Rommani
(Scandinavia) are usually referred to as Para-Romani
varieties.
In the absence of reliable census figures, the total
population of Romani speakers can only be estimated,
at anywhere upwards of 3.5 million. The largest concentrations of Romani speakers are in southeastern
and central Europe, especially Macedonia, Bulgaria,
Romania, and Slovakia. Romani has traditionally
been an oral language, and in more traditional communities there is even opposition to codification
attempts or other public use of the language, which
is viewed as having protective functions. The overwhelming trend, however, since the early 1990s has
been toward codification of the various dialects at
local or regional levels. The language is now used in
local media, on numerous Internet sites, as a medium
of correspondence (especially electronic), and in some
countries even as a medium of school instruction.
History
The earliest attestation of Romani is from 1542, in
Western Europe. Our understanding of the languages
historical development is therefore dependent on reconstruction and comparison with other Indo-Aryan
idioms as well as with the contact languages. In phonology, Romani shares a number of ancient isoglosses
with the Central branch of Indo-Aryan, most notably
the realization of Old Indo-Aryan r as u or i (San

of ks. - as kh
skrit srn. - Romani sun- to hear) and
(Sanskrit aks. i Romani j-akh eye). In contrast, however, to the other Central languages, Romani preserves
a number of dental clusters (Romani trin three,
phral brother; cf. Hindi tn, bhai). This led Turner
(1926) to assume a Central origin of Romani, with
subsequent migration to the Northwest before the
reduction of the relevant clusters took place. A northwestern migration is of course well in line with an
utlimate migration out of India and on towards
Europe. Further support for Turners theory comes
from the domain of verb morphology, where Romani
follows the exact same pattern as Northwestern languages such as Kashmiri or Shina in its renewal of
the past-tense conjugation through the adoption of
oblique enclitic pronouns as person markers (kerdo
done me me > kerdjom I did). Proto- or preEuropean Romani was thus a kind of Indian hybrid:
a central Indic dialect that had undergone partial
convergence with northern Indic languages. Although
the retention of dental clusters would suggest a break
with the Central languages during the transition
period from Old to Middle Indo-Aryan, the overall
morphology of Romani indicates that the language
participated in some of the significant developments
leading toward the emergence of New Indo-Aryan
(such as the reduction of the nominal case system to
a two-way opposition, nominative vs. oblique, and
grammaticalisation of new, postposed case markers).
It would appear therefore that Proto-Romani did not
leave the Indian subcontinent until late in the second
half of the first millennium CE. Romani is among
the most conservative New Indo-Aryan languages in
retaining a full consonantal present conjugation, as
well as consonantal oblique nominal case endings.
Typical phonological developments that characterize
Romani among the Indo-Aryan languages are the
devoicing of aspirates bh, dh, gh to ph, th, kh, the
shift of medial d, t to l, of short a to e, of inflectional
-a to -o, of initial kh to x, and of the retroflexes d. , .t,
d. d. , .t.t, d. h etc. to r and r.
The subsequent development of the language was
strongly influenced by its contact languages. Romani
borrowed lexicon and some grammatical vocabulary
from Iranian languages and Armenian. The heaviest
impact on Early Romani (European Romani, between
the 10th and 13th centuries C.E.) was of Byzantine
Greek. Apart from numerous lexical loans, phonemes,
and grammatical vocabulary, Romani adopted Greek
inflectional morphology in nouns and verbs, which
remain productive with loan vocabulary from subsequent European contact languages (see below). Greek
Romani 657
also had a strong impact on the syntax of Romani,

triggering among other things a shift to VO word order and the emergence of a preposed definite article.
The sound system

Romani dialects generally preserve an aspirated set
of voiceless stops ph, th, kh as well as c h, alongside
p, t, k, c and b, d, g, dz . Nasals are m and n, fricatives
are f, v, x, h, s, z, s , and in some dialects also z ,
and there is an affricate c [ts]. All dialects have l
and r, and some also retain r, which is realized as
either a uvular [R], a long trill [rr], or in some dialects a retroflex [8, ]. Palatalization of consonants,
either distinctive or nondistinctive, is common in the
Romani dialects of eastern and southeastern Europe.
The vowel system consists of a, e, i, o, u, with addition in some dialects of a central vowel e or i. Western
European dialects of Romani tend to show vowel
length distinctions. The phoneme inventory of individual dialects usually accommodates additional phonemes from the respective contact languages in lexical
loans. Conservative stress in Romani is on the final
inflectional segment of the word, though a number
of affixes remain unstressed, among them the vocative ending, agglutinative (Layer II) case endings (see
below), and the remoteness tense marker. Dialects in
Western and Central Europe often show a shift of
stress to earlier positions in the word.
Morphology
Nominal forms
Romani nominal morphology is inflectional, with

some agglutination. There are two genders,masculine
and feminine, and two numbers, singular and plural.
Mass nouns often allow omission of overt plural
marking. The principal inflectional alternation in
the noun is between two basic or Layer I cases,
nominative and oblique, in the singular and plural.
The different patterns of alternation constitute declension classes. Romani declension classes are sensitive to gender, to the phonological shape of the
stem, and to etymology, with European loans taking Greek-derived case endings. Basic inherited
declension classes are presented in Table 1.
Individual dialects show various patterns of analogies among the different classes. Loan declension
classes typically have Greek-derived inflection endings -os, -o, -is, or -us (masculine) and -a (feminine),
with a variety of plural endings such as -i, -e, -ides,
-uri and more. The oblique stem serves as the base
for further (Layer II) agglutinative case formation,
with the endings -te/-de (locative and prepositional),
Table 1 Basic ikeoclitic declension classes
Masculines
in o
Masculines
in f
Feminines
in i
Feminines
in f
Sg.
nominative
Sg.
oblique
Pl.
nominative
Pl.
oblique
chav-o
chav-es-
chav-e
chav-en-
phral-es-
phral-a
phral-en
romn-ja-
romn-ja
romn-jen-
phen-a-
phen-a
phen-en-
boy
phral
brother
romn-i
woman
phen
sister
-ke/-ge (dative), -tar/-dar (ablative), -sa(r) (instrumental and comitative), and -ker-/-ger- (genitive). As in
other Indo-Aryan languages, the genitive agrees with
the head noun (c hav-es-ker-o phral the boys brother, c hav-es-ker-i phen the boys sister). The oblique
without a Layer II extension serves as the case of the
direct object (accusative) with animate nouns.
Adjectives usually take vowel endings that agree
with the vocalic case-endings of the noun (mir-o dad
my father, mir-i daj my mother). Demonstratives
usually show a four-term system, encoding both proximity/remoteness (or, rather, presence in the situation
vs. the discourse context), and general/specific (disambiguation), e.g., adava, akava, odova, okova.
Interrogatives are cognate with other Indo-Aryan languages (kon who, kaj where), with so what
serving as the base for several derived forms (savo
which, soske why, sode how many, etc.). Indefinite markers are often borrowed from the respective
contact languages.
Verbs
Valency is a central feature of Romani verb morphology. It is expressed through direct affixation to the verb
root. The productivity, however, of individual valency
markers varies among the dialects. Typical valencyincreasing markers are -av-, -ar-, -ker-, and valencydecreasing markers are -jov- and -av-. They derive
verbs from other verb roots, as well as from nouns
and adjectives. Borrowed verbs carry loan verb extension or adaptation markers, based on Greek-derived
tense/aspect affixes such as -iz-, -in-, -is-, sometimes in
combination with valency affixes (e.g., -is-ar-, -is-ker-).
The default stem (root with derivation marker)
serves as a non-perfective aspect. The plain form
of the nonperfective serves as a present/subjunctive.
A tense/modality extension -a marks the present/
indicative, the future, or conditional, depending
on the dialect. A perfective aspect (also aorist or
simple past) is formed by attaching a perfective
extension (derived from the Middle Indo-Aryan
participle extension -t-) to the root of the verb
658 Romani
(e.g. ker-d-did). The choice of perfective extension

depends on the numerous perfective classes, which
are sensitive to the root phonology as well as to
valency and semantics.
There are two person conjugations: The present
conjugation (1sg -av, 2sg -es, 3sg -el, 1pl -as, 2/3pl
-en) continues the Middle Indo-Aryan set of present
concord markers. There are two inflection classes in
the present (nonperfective), distinguishing vocalic
and consonantal roots (xa-s you eat, kam-es youwant). The perfective conjugation, which follows the
perfective extension, derives from late Middle IndoAryan enclitic pronouns (1sg -om, 2sg -al/an, 3sg -as,
1pl -am, 2pl -an/-en, 3pl -e).
Both the present and the perfective may be extended by a remoteness marker -as/-ahi/-ys/-s that is external to the subject concord marker, indicating the
imperfect/habitual/conditional (with the present) or
the pluperfect/counterfactual (with the perfective).
Syntax
Romani stands out among the Indic languages
through its Europeanized, specifically Balkanized syntax. Word order is VO, with variation between thetic
(continuative) VS and categorical (contrastive) SV.
Local relations are indicated by prepositions. Adjectives and determiners generally precede the noun, as
does the definite article (which agrees with the noun
in gender, number, and case). Relative clauses are
postposed, and often introduced by a universal relativizer kaj < where. Clauses are generally finite.
Adverbial clauses are introduced by conjunctions,
usually derived from interrogatives. Romani distinguishes between factual and nonfactual complex
clauses. Modal, manipulation, and purpose clauses
are introduced by a nonfactual conjunction te, as are
conditional clauses. Epistemic complements are introduced by kaj, which is often replaced by a borrowing
from the respective contact language.
Dialect diversity
Dialect differentiation in Romani appears to have
emerged largely in situ, following the dispersal of
groups from the Balkans into western and northern
Europe, from around the 14th century onward, and
their settlement in their present locations, during the
16th-17th centuries. There are two major diffusion
centers of innovations: in the southeast, especially
the northern Balkans, and in western-central Europe,
especially Germany. Typical of the western-northern
dialects are prothesis of j-, simplification of ndr to r,
loss of adjectival past-tense in intransitives (gelo, geli
> geljas he/she went), and retention of -n in the
abstract nominalizer -ipen/-iben. In the central
regions, s in grammatical paradigms is often replaced
by h. Individual regions show distinct developments
in morphological paradigms, especially demonstratives, 2/3pl perfective concord markers, and loan
verb markers. Especially these latter isoglosses justify
the current classification into the following dialect
groups: Balkan (with a subgroup Black Sea Coast),
Vlax (Transylvannia and adjoining regions), Central,
Northeast (Baltic-Northrussian), and Northwest
(German-Scandinavian).
See also: Aelfric (fl. 987-1010); Armenian; Balkans as a
Linguistic Area; Domari; Dravidian Languages; Hindi;
IndoAryan Languages; Iranian Languages; Kashmiri.
Bibliography
Bakker P & Matras Y (2003). Bibliography of modern
Romani linguistics. Amsterdam: John Benjamins.
Els k V & Matras Y (eds.) (2000). Grammatical relations in
Romani: the noun phrase. Amsterdam: John Benjamins.
Matras Y (2002). Romani: a linguistic introduction.
Turner R (1926). The position of Romani in Indo-Aryan.
Journal of the Gypsy Lore Society 5(3), 145189.
Romania: Language Situation

D R Ladd, The University of Edinburgh, Edinburgh, UK
The central feature of the language situation in

Romania, as in several other countries of Central
and Eastern Europe, such as Bulgaria and Poland, is
the existence of stable minority-language communities within the geographical confines of what
would otherwise be a relatively homogeneous nation

state. Any description of the situation must take into
account the historical tensions among the various
ethnic or national groups, the post-war Communist
governments efforts to create a strong centralized
administration, and the still uncertain political future.
The broad outlines of the situation are straightforward. The only official language is Romanian, spoken
as a first language by approximately 88% of the
658 Romani
(e.g. ker-d-did). The choice of perfective extension

depends on the numerous perfective classes, which
are sensitive to the root phonology as well as to
valency and semantics.
There are two person conjugations: The present
conjugation (1sg -av, 2sg -es, 3sg -el, 1pl -as, 2/3pl
-en) continues the Middle Indo-Aryan set of present
concord markers. There are two inflection classes in
the present (nonperfective), distinguishing vocalic
and consonantal roots (xa-s you eat, kam-es youwant). The perfective conjugation, which follows the
perfective extension, derives from late Middle IndoAryan enclitic pronouns (1sg -om, 2sg -al/an, 3sg -as,
1pl -am, 2pl -an/-en, 3pl -e).
Both the present and the perfective may be extended by a remoteness marker -as/-ahi/-ys/-s that is external to the subject concord marker, indicating the
imperfect/habitual/conditional (with the present) or
the pluperfect/counterfactual (with the perfective).
Syntax
Romani stands out among the Indic languages
through its Europeanized, specifically Balkanized syntax. Word order is VO, with variation between thetic
(continuative) VS and categorical (contrastive) SV.
Local relations are indicated by prepositions. Adjectives and determiners generally precede the noun, as
does the definite article (which agrees with the noun
in gender, number, and case). Relative clauses are
postposed, and often introduced by a universal relativizer kaj < where. Clauses are generally finite.
Adverbial clauses are introduced by conjunctions,
usually derived from interrogatives. Romani distinguishes between factual and nonfactual complex
clauses. Modal, manipulation, and purpose clauses
are introduced by a nonfactual conjunction te, as are
conditional clauses. Epistemic complements are introduced by kaj, which is often replaced by a borrowing
from the respective contact language.
Dialect diversity
Dialect differentiation in Romani appears to have
emerged largely in situ, following the dispersal of
groups from the Balkans into western and northern
Europe, from around the 14th century onward, and
their settlement in their present locations, during the
16th-17th centuries. There are two major diffusion
centers of innovations: in the southeast, especially
the northern Balkans, and in western-central Europe,
especially Germany. Typical of the western-northern
dialects are prothesis of j-, simplification of ndr to r,
loss of adjectival past-tense in intransitives (gelo, geli
> geljas he/she went), and retention of -n in the
abstract nominalizer -ipen/-iben. In the central
regions, s in grammatical paradigms is often replaced
by h. Individual regions show distinct developments
in morphological paradigms, especially demonstratives, 2/3pl perfective concord markers, and loan
verb markers. Especially these latter isoglosses justify
the current classification into the following dialect
groups: Balkan (with a subgroup Black Sea Coast),
Vlax (Transylvannia and adjoining regions), Central,
Northeast (Baltic-Northrussian), and Northwest
(German-Scandinavian).
See also: Aelfric (fl. 987-1010); Armenian; Balkans as a
Linguistic Area; Domari; Dravidian Languages; Hindi;
IndoAryan Languages; Iranian Languages; Kashmiri.
Bibliography
Bakker P & Matras Y (2003). Bibliography of modern
Romani linguistics. Amsterdam: John Benjamins.
Elsk V & Matras Y (eds.) (2000). Grammatical relations in
Romani: the noun phrase. Amsterdam: John Benjamins.
Matras Y (2002). Romani: a linguistic introduction.
Turner R (1926). The position of Romani in Indo-Aryan.
Journal of the Gypsy Lore Society 5(3), 145189.
Romania: Language Situation

D R Ladd, The University of Edinburgh, Edinburgh, UK
The central feature of the language situation in

Romania, as in several other countries of Central
and Eastern Europe, such as Bulgaria and Poland, is
the existence of stable minority-language communities within the geographical confines of what
would otherwise be a relatively homogeneous nation

state. Any description of the situation must take into
account the historical tensions among the various
ethnic or national groups, the post-war Communist
governments efforts to create a strong centralized
administration, and the still uncertain political future.
The broad outlines of the situation are straightforward. The only official language is Romanian, spoken
as a first language by approximately 88% of the
Romania: Language Situation 659
population (1977 census figures). The standard language is essentially uniform throughout the country,
having been established as a standard only in the 19th
century, though there is a fair amount of variation in
nonstandard rural speech. Standard Romanian is the
language of administration and virtually all higher education. During the Communist period, members of
the national minorities (referred to by the Romanian
phrase popoarele conlocuitoare, lit. co-inhabiting peoples) were expected to learn Romanian in school, and
virtually all did so. However, there were provisions for
the extensive use of minority languages in primary and
secondary education, publishing, broadcasting, and
the arts provisions comparable to those made by
other Communist governments in post-World War II
Eastern Europe. Wherever the ethnic composition of
the population warranted it, newspapers and books
were published in minority languages, and schools
and state-supported theatres in minority languages
were established.
Romanias current territory is traditionally divided
into three provinces: Wallachia (Rom. ara Roma neasca ) to the south, Moldavia (Rom. Moldova) to
the east, and Transylvania (Rom. Transilvania or
Ardeal, Hung. Erde ly, Ger. Siebenbu rgen) to the north
and west on the continental side of the Carpathian
Mountains. The two main groups of minority language speakers, Hungarians and Germans, are concentrated in Transylvania, which has a long history
as a semi-independent feudal state. As early as the
14th century, the Transylvanian nobility was almost
exclusively Hungarian, whereas the military and merchant classes in the developing cities were in many
cases German. The peasantry was by no means
entirely Romanian, but most Romanians were peasants. Romanian nationalist movements were active during the late 18th and 19th centuries, when
Transylvania was part of the Austro-Hungarian
Empire, and in 1920, the province was awarded to
Romania under the Treaty of Trianon. The catalogue
of wrongs arising from this complex history is both
substantial and tended carefully by all the groups
concerned.
There are Hungarians living throughout Transylvania. They represent some 8% of the population of
Romania as whole, but in certain areas of the province, they constitute a substantial majority. They do
not regard themselves as Romanian. Hungarian
remains in active use almost everywhere in the
officially supported contexts discussed above, and in
shops and on unofficial (but not official) signs. The

Germans, by contrast, are restricted to scattered settlements in and around the cities in the south of the
province (e.g., Sibiu, Ger. Hermannstadt) and various
rural communities, especially in the Banat (the far
western area around Timzis oara). At the time of
the 1977 census, the Germans made up 1.2% of the
population of Romania as a whole, but under
the terms of an agreement with West Germany, a
substantial proportion of the German population
emigrated during the 1980s, and it is not clear whether the German-speaking communities will survive
for much longer.
Throughout the country, there are small populations
of speakers of other languages, including Serbian,
Ukrainian, and Romany (Romani). Together, they
constituted 2.8% of the population in 1977. In addition, there is a large population of Romanian speakers
in the former Soviet Republic of Moldavia (the eastern half of the historical province of Moldavia), now
generally known in English by its Romanian name
Moldova. During the Soviet period, Moldavian
was written in the Cyrillic alphabet and treated as a
separate language. The orthography has since
reverted to that of standard Romanian.
Toward the end of the Ceaus escu regime, many of
the provisions for minority languages were dismantled, in what members of the minority communities
saw as deliberate Romanianization. This perception
may be exaggerated: Ceaus escus policies (in particular, the destruction of villages and the systematization of their displaced inhabitants in agro-industrial
complexes) often affected ethnic Romanians as well.
Nevertheless, the Romanian nationalist bent to
Ceaus escus policies was unmistakable, and nationalism remains a powerful force in post-Communist
politics. It is difficult to predict how it will affect the
language situation in the coming decades.
See also: Hungary: Language Situation; Moldova: Language Situation.
Bibliography
Grimes B (ed.) (2002). The ethnologue (14th edn.). Dallas:
SIL International.
Petyt K M (1975). Romania a multilingual nation. International Journal of the Sociology of Language 4, 75101.
Posner R (1996). The Romance languages. Cambridge:
660 Romanian
Romanian
J Augerot, University of Washington, Seattle, WA, USA
The Romanian (alternatively spelled Rumanian,

Roumanian) language is a member of the Romance
languages and is a continuation of Eastern Latin, spoken in the Roman province of Dacia and surrounding
areas which were colonized by the Romans led by the
emperor Trajan from 106 A.D.
The Eastern Latin spoken in this area by Roman
soldiers and colonists along with presumably assimilated local tribes underwent numerous influences after
the Roman legions abandoned the northernmost areas
in 271 A.D. The first major influence was undoubtedly
provided by the indigenous DacianThracianIllyrian
inhabitants of the area, although only a few undisputed words from that source remain. A. Rosetti mentions
some 88 Indo-European but pre-Latin terms whose
existence in both Romanian and Albanian suggests a
substratum as the source for these words in both languages, e.g., barza stork, bucurie happiness, ma nz
colt, mos old man; cf. Albanian bardhe white
(FEM), bukuri beauty, me z colt, moshe (old) age.
It is likely that some of Romanians morphosyntactic structure, especially that shared with the nonRomance Balkan languages, also comes from this
pre-Latin substratum, but the complete lack of textual evidence leaves these possibilities in the realm of
speculation.
The longest and most consistent influence was that
of the Slavs, who began to migrate across this territory to the south as early as 400 A.D. and who remain
neighbors on three sides of present-day Romania.
Several other migrations of peoples crossed the territory of todays Romania between the departure of the
Roman legions and the appearance of the modern
Romanian people, but little or no linguistic traces of
the Goths, Huns, Gepids, Avars, et al. remain.
This early Romanian soon (perhaps as early as the
10th century) began to split, first into four dialects
which later tended to become languages in their own
right: the principal one in terms of numbers is
termed Daco-Romanian, spoken primarily north of
the Danube. The second largest division is called
Aromanian (Macedo Romanian) and is currently
spoken in pockets of southwestern Bulgaria, Macedonia, Albania, and northern Greece. The other
two are quite limited in extent: Megleno Romanian
is confined to an area in southeastern Macedonia
and adjacent northern Greece, and Istro Romanian
is limited to eight localities in Istria in modern
Croatia.
Daco-Romanian is now called Romanian and is

the single national language among the four with
approximately 25 million speakers worldwide. It
shows, in addition to the Slavic mentioned above,
the influences of prolonged contact with Hungarian,
Turkish, and early modern Greek. More recently,
from the 19th century onwards, Romanian has been
subjected to a kind of re-Latinization, both from the
strong influence of French and from the international
European vocabulary largely based on Latin and
Italian.
The first preserved texts in Romanian date from
the end of the 15th and beginning of the 16th centuries. The oldest dated text (1521) is a letter to Neacsu
of Ca mpulung.
Although some texts from as early as the 16th
century are written in the Latin alphabet, which is
the norm today, until 1860 the official alphabet was
Cyrillic. Most of the letters have approximately their
general European values, but there is some use of
diacritics that should be noted, namely , a , a , t, s .
Phonetic equivalents are given where necessary in
brackets in Tables 1 and 2.
Romanian symbolizes diphthongs with pairs of
vowels, e.g., iu, ia, ea, ua, oa and ai, au, ou, ui, but
one can seldom be certain from the spelling alone. See
the website of the Romanian Academy Center for
Artificial Intelligence for a complete inventory of
multiple vowel combinations.
In contrast to the practice of some of its immediate
neighbors, there is no final devoicing of consonants as
Table 1 Vowels
High
Mid
Low
Front
Central
Back rounded
i
e
(a) [i$]
a [e]
a
u
o
Table 2 Consonants
Bilabial
Stops
Affricates
Fricatives
Nasals
Lateral
Trill
Labiodental
pb
fv
m
Dental
Palatoalveolar
Dorsovelar
td
ch [k] gh
[g]
c [c] g [ ]
[s] j [z]
c [k] g
[c]
sz
n
l
r
Romanian 661
Table 3 Morphophonemics
underlying forms
mes- mes-e
fet- fet-e
mer-u mer-e
per-u per-i
1. breaking
2. backing
3. coalescence
4. drop final /u, i/
me9ase
me9ase
mase mese

fe9ate
feate
fate fete

meru

mer mere

peru

per per
normal spelling
masa mese
fata fete
mar mere
par peri
in Bulgarian and Russian and no contrastive vowel

length as in Slovak, Hungarian, and Serbian (SerboCroatian). Rather, Romanian forms a bridge between Serbian and Ukrainian in terms of final voiced
obstruents.
The two palatal velars are disputed as separate
from the dorsovelars in phonology but are clearly
different on the morphophonemic plane. They represent an interesting progression from Latin cl and gl
in words such as cheie key and gheata ice. Intentionally omitted from the chart are the palatalized
variants of most consonants that, according to some
analyses, occur word finally before the final (voiceless?) [i], for example in the second person singular
present forms of the verb, e.g., tu plimbi you stroll,
vezi you see, scoli you arise.
Word stress in Romanian is basically penultimate,
but only if one counts final (mostly silent and, in the
case of /u/, mostly unwritten) /i/, /u/ as a syllable (the
stressed vowel is given in bold type): plecare-pleaca plecand, venire-vine-venind, miere. There are several
exceptions to this generalization including many
borrowings, e.g., through Turkish: baklava, sarma,
cafea, and the majority of the infinitive forms, which
are also stressed on the final syllable: veni, citi, pleca,
turna, putea (historically, of course, the infinitive had
a final syllable -re). Other exceptions include the
classic Latin deviations based on heavy syllables as
in industrie and words of specific morphological categories with stress fixed on a certain ending, such as the
imperfect tense: plecam, plecai, pleca, etc. Note that
the addition of the masculine genitive definite article
/-u-lui/, the feminine definite plural article /-i-le/ or
the multisyllabic plural desinence /-uri/ does not
cause movement of the stress: marfa , marfuri. Thus
it is not truly penultimate, only dominantly so, since
one must proceed from a hypothetical base form of
the word and include several classes of exceptions.
From a morphophonemic point of view, Romanian
is quite complex. Space limitations will not allow a
complete treatment here, but an illustration of this
complexity may be seen in the singular and plural
forms of words such as masa , mese table(s); fata ,
fete girl(s); ma r, mere apple(s); pa r, peri hair(s),
Table 4 Verbal forms with root in /e/

present 1st sing.
present 2nd sing.
present 3rd sing.
conjunctive 3rd sing.
vad
vezi
vede
sa vada
nva
nve i
nva a
sa nve e
where only by beginning with a root with the front

vowel /e/ may one predict all the forms of any one
word. The three rules that are involved are illustrated
in Table 3.
Rule 4 brings up some other complications such as
the phonetic interpretation of consonants preceding
the final dropped /-i/, but it can be seen that, at least
for inherited Latin vocabulary, these rules will predict
the surface phonetics of many forms. Compare (given
here in the normal orthography) the various forms of
the verbs a vedea to see and a nvata to learn with
the same stressed root vowel /e/, i.e. /ve d-/, /i$nve c-/
(see Table 4).
It can be seen that the same rules allow the derivation of all the verbal forms if we add the flip-flop
rule, namely, in the conjunctive-subjunctive the present third singular ending /e/ goes to /e/ and vice
versa. Thus el pleaca he leaves becomes el o sa plece
while el crede he believes becomes el o sa creada .
It should be noted that the stressed schwa is a
feature that is often included among the features of
the so-called Balkan Sprachbund. There are characteristics (merging of the genitive and dative, tendency
to lose the infinitive, development of a postpositive
definite article, repetition pronominally of the direct
object) in several areas of the grammar that Romanian
shares with languages clearly not of Romance origin such as Bulgarian, Macedonian, Albanian, and
dialectal areas of Serbian and modern Greek.
Most noticeable in differentiating Romanian from
the rest of Romance is the postposed definite article,
which takes several forms but is basically singular -a,
plural -le for feminines and singular -l, plural -i for
masculines (with a special category of neuter that
generally takes the masculine article in the singular
and the feminine in the plural; examples in Table 5).
662 Romanian
In the noun phrase the same definite article is attached

to the first element in the phrase: un prim pas a first
step, primul pas the first step.
Unlike the rest of Romance, although they have
been whittled down to just two (plus a vocative that
has almost disappeared), Romanian still maintains
nominal case distinctions. Among Romanian scholars there is a tendency to name them nominativeaccusative and genitive-dative after the functions of
each case; see Table 6 for some examples.
Table 7 gives two verbs illustrating the basic verbal
categories: a intra to enter and a face to do. There
are, in addition to the present tense, two ways to form
the future, and a variety of past tenses, including
the compound past, the imperfect, the perfect and
the pluperfect. Again it is instructive to consider the
dropped final high vowels /u/, /i/ as morphological
formants of the first and second persons singular in
Table 5 The definite article

Feminine
Masculine
Neuter
o doamna a lady
doamna doamnele
o masa a table
masa mesele
o carte a book
cartea car ile
un domn a gentleman
domnul domnii
un pom a tree
pomul pomii
un perete a wall
peretele pere ii
un os a bone
osul oasele
un lucru a thing
lucruri lucrurile
un scaun a chair
scaunul scaunele
the present tense, both for purposes of word accent

and for various other phonological alternations. The
forms are given in standard orthography.
There are two general types of regular verb, one
with base back vowel and the other with base
plus front vowel, and with this information plus
sets of personal endings all of the conjugations of a
given verb may be produced. I chose here the verb a
intra since it gives justification for the otherwise
phantom ending /-u/ of the first person singular for
a-verbs and both first singular and third plural for
e-verbs (also true for i-verbs, which are not illustrated
here).
All other verbal formations (except for the
conjunctive-subjunctive future with o sa plus the
present, discussed above with the flip-flop rule) are
synthetic, relying on either the short infinitive or the
past passive participle: thus the future is voi, vei, va,
vom, veti, vor plus the infinitive (voi intra, vei intra,
etc.), and the conditional is as , ai, ar, am, ati, ar also
with the infinitive (as intra, ai intra, etc.); while the
compound past uses am, ai, a, am, ati, au plus the past
passive participial form (am intrat, ai intrat, etc.), and
the past conditional also uses it after the same auxiliary as the conditional, adding the verb fi to be (as
fi intrat, ai fi intrat, etc.). One further verb form is the
past subjunctive, which does not conjugate and has a
single fixed auxiliary sa fi with the past passive participle (sa fi intrat for all persons). There is also a
future in the past with fi.
Table 6 Cases
Gender
Article
Nom-Acc Singular
Gen-Dat Singular
Nom-Acc Plural
Gen-Dat Plural
Feminine
Indefinite
Definite
Indefinite
Definite
Indefinite
Definite
casa house
casa
barbat man
barbatul
lemn wood
lemnul
case
casei
barbat
barbatului
lemn
lemnului
case
casele
barba i
barba ii
lemne
lemnele
case
caselor
barba i
barba ilor
lemne
lemnelor
Masculine
Neuter
Table 7 Verb tenses
Present
Imperfect
Simple perfect
Pluperfect
1st sing
2nd
3rd
1st plur
2nd
3rd
intru
fac
intram
faceam
intrai
facui
intrasem
facusem
intri
faci
intrai
faceai
intra i
facu i
intrase i
facuse i
intra
face
intra
facea
intra
facu
intrase
facuse
intram
facem
intram
faceam
intraram
facuram
intraseram
facuseram
intra i
face i
intra i
facea i
intrara i
facura i
intrasera i
facusera i
intra
fac
intrau
faceau
intrara
facura
intrasera
facusera
Romanian Lexicography 663

See also: Balkans as a Linguistic Area; Romania: Lan-
guage Situation; Romanian Lexicography.
Bibliography
Agard F (1958). Structural sketch of Rumanian. Baltimore:
Waverly Press.
Augerot J E (1974). Romanian phonology. Moscow, ID:
Idaho Research Foundation.
Alexandrescu S (ed.) (1977). Transformational grammar
and the Rumanian language. Lisse, Netherlands: The
Peter de Ritter Press.
Chitoran I (2002). The phonology of Romanian: a constraint-based approach. New York: Mouton de Gruyter.
Coteanu I & Danaila I (1970). Introducere n lingvistica s i
filologia roma neasca . Bucharest: Editura Academiei
Republicii Socialiste Romania.
Petrovici E (1934). Le pseudo i final du roumain. Bulletin
linguistique 2, 86.
Pop L & Moldovan V Grammaire du roumain/Romanian
grammar. Cluj: Echinox.
Puscariu S (1940). Limba romana . Bucharest: Funda ia

Pentru Literatura si Arta.
Rosetti A (19641966). Istoria limbii roma ne (6 vols).
Bucharest: Editura Stiin ifica.
Rosetti A (1973). Bre`ve histoire de la langue roumaine des
origines a` nos jours. The Hague: Mouton.
Sala M (1976). Contributions a` la phone tique historique du
roumain. Paris: Klincksiek.
Stan I (1996). Studii de fonetica s i fonologie. Cluj-Napoca:
Presa Universitara Clujeana.
Vasiliu E (1965). Fonologia limbii romane. Bucharest: Editura
Stiin ifica.
Vasiliu E & Golopentia Eretescu S (1972). The transformational syntax of Romanian. The Hague: Mouton.
Relevant Website
http://www.racai.ro Romanian Academy Center for
Artificial Intelligence.
Romanian Lexicography
D Tufis, Romanian Academy Institute for Artificial
Intelligence, Bucharest, Romania
Romanian lexicography has its roots in the 16th

century, when the first Slavic-Romanian dictionaries
where compiled, mainly to support translation of
religious texts. The following century, Slavic was increasingly replaced by Latin or Romance languages
(especially Italian) and by 1700 the first RomanianLatin dictionary was compiled, Anonymus Caransebiensis (containing more than 5,000 entries). At
first, Romanian dictionaries had circulated only as
manuscripts. The first printed dictionary (four languages) appeared in 1825Lexiconul de la Buda
(Lesiconul roma nesc-latinesc-unguresc-nem esc)
as a joint effort of the Scoala Ardeleana scholars.
This is considered to be the starting point for the
modern Romanian lexicography. In 1870 and 1879
Al. Cihac published (in French) the two volumes of
his etymological dictionary Dic ionarul etimologic
dacoroma n.
The Romanian Academy, founded in April 1866,
developed plans for a large dictionary of Romanian
as complete and rational as possible, to become a
guide to the language (Sala and Danaila, 1997).
A first attempt was made in 1869, when A. T. Laurian
and I. C. Massim started work on Dic ionarul limbei
roma ne, containing words of Latin origin, and Glossariul, containing words of foreign form and of uncertain origin. The two volumes, with 50 000 entries,
appeared in 1873 and 1877 respectively, but were
roundly criticized for excessive Latinizing of the
Romanian language. However, this dictionary has the
undeniable merit of being among the first dictionaries
to use the principle of multiple etymologies. In 1884
editorship of the dictionary passed to B. P. Hasdeu,
who produced three volumes (1887 and 1893) of the
well-known Etymologicum magnum Romaniae, covering letters A and B (up to ba rbat). Continuation of
the task was assigned in 1897 to A. I. Philippide who,
with his team, completed the dictionary up to the
word de. When the Dictionary Commission of the
Romanian Academy asked for the removal of several
neologisms, Phillippide gave up the task (1905) and
his partial dictionary was never published.
The Dic ionarul limbii roma ne/Dic ionarul Academiei (DLR/DA) in its present form, was started by
S. Puscariu. He and his team in Cluj worked for 43
years and published, in several fascicles, the letters A,
B, C, D (up to de) F, G, H, I, I , J, and L (up to lojni a ),
altogether 3075 pages, containing about 60 000
words and variants. Puscariu died in 1948, but his
work was continued in Bucharest at the Institute of
Linguistics, headed by Iorgu Iordan. Unfortunately,
the work was interrupted in 1952 on the grounds that
a general dictionary for use by ordinary people was
Romanian Lexicography 663

See also: Balkans as a Linguistic Area; Romania: Lan-
guage Situation; Romanian Lexicography.
Bibliography
Agard F (1958). Structural sketch of Rumanian. Baltimore:
Waverly Press.
Augerot J E (1974). Romanian phonology. Moscow, ID:
Idaho Research Foundation.
Alexandrescu S (ed.) (1977). Transformational grammar
and the Rumanian language. Lisse, Netherlands: The
Peter de Ritter Press.
Chitoran I (2002). The phonology of Romanian: a constraint-based approach. New York: Mouton de Gruyter.
Coteanu I & Danaila I (1970). Introducere n lingvistica si
filologia romaneasca. Bucharest: Editura Academiei
Republicii Socialiste Romania.
Petrovici E (1934). Le pseudo i final du roumain. Bulletin
linguistique 2, 86.
Pop L & Moldovan V Grammaire du roumain/Romanian
grammar. Cluj: Echinox.
Puscariu S (1940). Limba romana. Bucharest: Funda ia

Pentru Literatura si Arta.
Rosetti A (19641966). Istoria limbii romane (6 vols).
Bucharest: Editura Stiin ifica.
Rosetti A (1973). Bre`ve histoire de la langue roumaine des
origines a` nos jours. The Hague: Mouton.
Sala M (1976). Contributions a` la phonetique historique du
roumain. Paris: Klincksiek.
Stan I (1996). Studii de fonetica si fonologie. Cluj-Napoca:
Presa Universitara Clujeana.
Vasiliu E (1965). Fonologia limbii romane. Bucharest: Editura
Stiin ifica.
Vasiliu E & Golopentia Eretescu S (1972). The transformational syntax of Romanian. The Hague: Mouton.
Relevant Website
http://www.racai.ro Romanian Academy Center for
Artificial Intelligence.
Romanian Lexicography
D Tufis, Romanian Academy Institute for Artificial
Intelligence, Bucharest, Romania
Romanian lexicography has its roots in the 16th

century, when the first Slavic-Romanian dictionaries
where compiled, mainly to support translation of
religious texts. The following century, Slavic was increasingly replaced by Latin or Romance languages
(especially Italian) and by 1700 the first RomanianLatin dictionary was compiled, Anonymus Caransebiensis (containing more than 5,000 entries). At
first, Romanian dictionaries had circulated only as
manuscripts. The first printed dictionary (four languages) appeared in 1825Lexiconul de la Buda
(Lesiconul romanesc-latinesc-unguresc-nem esc)
as a joint effort of the Scoala Ardeleana scholars.
This is considered to be the starting point for the
modern Romanian lexicography. In 1870 and 1879
Al. Cihac published (in French) the two volumes of
his etymological dictionary Dic ionarul etimologic
dacoroman.
The Romanian Academy, founded in April 1866,
developed plans for a large dictionary of Romanian
as complete and rational as possible, to become a
guide to the language (Sala and Danaila, 1997).
A first attempt was made in 1869, when A. T. Laurian
and I. C. Massim started work on Dic ionarul limbei
romane, containing words of Latin origin, and Glossariul, containing words of foreign form and of uncertain origin. The two volumes, with 50 000 entries,
appeared in 1873 and 1877 respectively, but were
roundly criticized for excessive Latinizing of the
Romanian language. However, this dictionary has the
undeniable merit of being among the first dictionaries
to use the principle of multiple etymologies. In 1884
editorship of the dictionary passed to B. P. Hasdeu,
who produced three volumes (1887 and 1893) of the
well-known Etymologicum magnum Romaniae, covering letters A and B (up to barbat). Continuation of
the task was assigned in 1897 to A. I. Philippide who,
with his team, completed the dictionary up to the
word de. When the Dictionary Commission of the
Romanian Academy asked for the removal of several
neologisms, Phillippide gave up the task (1905) and
his partial dictionary was never published.
The Dic ionarul limbii romane/Dic ionarul Academiei (DLR/DA) in its present form, was started by
S. Puscariu. He and his team in Cluj worked for 43
years and published, in several fascicles, the letters A,
B, C, D (up to de) F, G, H, I, I, J, and L (up to lojni a),
altogether 3075 pages, containing about 60 000
words and variants. Puscariu died in 1948, but his
work was continued in Bucharest at the Institute of
Linguistics, headed by Iorgu Iordan. Unfortunately,
the work was interrupted in 1952 on the grounds that
a general dictionary for use by ordinary people was
664 Romanian Lexicography
much more important. Dic ionarul limbii roma ne literare contemporane (DLRLC) was intended to
answer this demand and under the coordination of
D. Macrea and E. Petrovici it was compiled, based
on DLR/DA, and published in 19551957. The four
volume dictionary, containing 48 000 headwords,
hardly achieved its original goal of being an ordinary
peoples dictionary. A shorter version (all citations
removed, but about 1500 new entries added) was
published in 1958 as Dic ionarul limbii roma ne
moderne (DM).
DLRLC and DM were superseded by the
Dic ionarul explicativ al limbii roma ne (DEX) published in 1975, followed in 1988 by Supliment la
Dic ionarul explicativ al limbii roma ne (DEX-S).
This dictionary, containing 56 569 entries (words
and variants) is the bestselling Dictionary of Romanian (Vintila -Ra dulescu, 2002). The second edition
(1996), republished in 1998, eliminated some inconsistencies, corrected several definitions, and added
almost 9000 entries. Since August 2004 it has been
publicly available on line.
Compilation of the most representative Romanian
lexicographic work, represented by the DLR/DA,
was resumed in 1959 at the Linguistic Institutes
in Bucharest, Cluj, and Ias i and the new series of
DLR fascicles began to be published in 1965. By
2004 32 volumes had been published, listing more
than 175 000 headwords and variants, with more
than 625 000 meanings and about 3 200 000 citations. Between 2001 and 2004 DLR has been published in an updated and shortened version (only four
volumes) as Micul Dic ionar Academic (MDA). To
meet the size restriction, MDA discarded the citations
and sources, and used a very compact encoding of
lexical information.
Romanian lexicography has produced many other

types of dictionaries, both monolingual and multilingual: dictionaries of neologisms, orthography, pronunciation and morphology, etymology, synonyms,
antonyms, terminology, encyclopedic dictionaries,
and regional and dialectal dictionaries.
See also: Academies: Dictionaries and Standards; Etymology; Lexicography: Overview; Romania: Language Situation; Romanian; WordNet(s).
Bibliography
Avram M, Sala M & Vintila -Ra dulescu I (1999). Institutul
de Lingvistica Iorgu Iordan din Bucures ti. 50 de ani
de existen a (19491999). Bucures ti: Editura Univers
Enciclopedic.
Sala M & Da naila I (1997). Lexicography at the Iorgu
Iordan Institute of Linguistics. In Tufis D & Andersen
P (eds.) Recent advances in Romanian language technology. Bucharest: Romanian Academy Publishing House.
9398.
Seche M (19661969). Schi a de istorie a lexicografiei
roma ne (2 vols). Bucures ti: Editura S tiintifica .
Vintila -Ra dulescu I (2002). Resurse lingvistice

pentru
limba roma na elaborate la Institutul de Lingvistica
Iorgu Iordan. In Tufis D & Filip Fl Gh (eds.) Limba
Roma na n Societatea Informa ionala Societatea
Cunoas terii. Bucures ti: Editura Expert. 1932.
Relevant Website
http://dexonline.ro The second edition of DEX.
Ronjat, Jules (18641925)

M Piot, University of Grenoble, Grenoble, France
Jules Ronjat, an amateur scholar, was born in 1864 in

Dauphine (France). He was closely bound up with
Provenc al authors such as the Nobel Literature
Prize-winner Fre de ric Mistral, whom he considered
the greatest authority on the Occitan language, and
the other Felibrige poets, as well as his Romanist
colleagues Camille Chabanneau, Emil Levy, Albert
Dauzat, etc. His work was influenced by French
Romanist linguists such as Gaston Paris, Father
Pierre-Jean Rousselot, and Maurice Grammont. He

earned his Doctorat-e`s-Lettres in 1913 from the Sorbonne for his theses Essai de syntaxe des parlers
provenc aux modernes and Le de veloppement du langage observe chez un enfant bilingue. In his later years
he taught as a privat-docent at the University of
Geneva.
In his reviews in the Revue des langues romanes, he
made, as did Grammont, a critical appraisal of the
linguistic geography practiced by Jules Gillie ron and
his followers. As denoted by the title (istorique) of
his first thesis, Essai de syntaxe des parlers provenc aux modernes, he was a champion of spelling reform
664 Romanian Lexicography
much more important. Dic ionarul limbii romane literare contemporane (DLRLC) was intended to
answer this demand and under the coordination of
D. Macrea and E. Petrovici it was compiled, based
on DLR/DA, and published in 19551957. The four
volume dictionary, containing 48 000 headwords,
hardly achieved its original goal of being an ordinary
peoples dictionary. A shorter version (all citations
removed, but about 1500 new entries added) was
published in 1958 as Dic ionarul limbii romane
moderne (DM).
DLRLC and DM were superseded by the
Dic ionarul explicativ al limbii romane (DEX) published in 1975, followed in 1988 by Supliment la
Dic ionarul explicativ al limbii romane (DEX-S).
This dictionary, containing 56 569 entries (words
and variants) is the bestselling Dictionary of Romanian (Vintila-Radulescu, 2002). The second edition
(1996), republished in 1998, eliminated some inconsistencies, corrected several definitions, and added
almost 9000 entries. Since August 2004 it has been
publicly available on line.
Compilation of the most representative Romanian
lexicographic work, represented by the DLR/DA,
was resumed in 1959 at the Linguistic Institutes
in Bucharest, Cluj, and Iasi and the new series of
DLR fascicles began to be published in 1965. By
2004 32 volumes had been published, listing more
than 175 000 headwords and variants, with more
than 625 000 meanings and about 3 200 000 citations. Between 2001 and 2004 DLR has been published in an updated and shortened version (only four
volumes) as Micul Dic ionar Academic (MDA). To
meet the size restriction, MDA discarded the citations
and sources, and used a very compact encoding of
lexical information.
Romanian lexicography has produced many other

types of dictionaries, both monolingual and multilingual: dictionaries of neologisms, orthography, pronunciation and morphology, etymology, synonyms,
antonyms, terminology, encyclopedic dictionaries,
and regional and dialectal dictionaries.
See also: Academies: Dictionaries and Standards; Etymology; Lexicography: Overview; Romania: Language Situation; Romanian; WordNet(s).
Bibliography
Avram M, Sala M & Vintila-Radulescu I (1999). Institutul
de Lingvistica Iorgu Iordan din Bucuresti. 50 de ani
de existen a (19491999). Bucuresti: Editura Univers
Enciclopedic.
Sala M & Danaila I (1997). Lexicography at the Iorgu
Iordan Institute of Linguistics. In Tufis D & Andersen
P (eds.) Recent advances in Romanian language technology. Bucharest: Romanian Academy Publishing House.
9398.
Seche M (19661969). Schi a de istorie a lexicografiei
romane (2 vols). Bucuresti: Editura Stiintifica.
Vintila-Radulescu I (2002). Resurse lingvistice

pentru
limba romana elaborate la Institutul de Lingvistica
Iorgu Iordan. In Tufis D & Filip Fl Gh (eds.) Limba
Romana n Societatea Informa ionala Societatea
Cunoasterii. Bucuresti: Editura Expert. 1932.
Relevant Website
http://dexonline.ro The second edition of DEX.
Ronjat, Jules (18641925)

M Piot, University of Grenoble, Grenoble, France
Jules Ronjat, an amateur scholar, was born in 1864 in

Dauphine (France). He was closely bound up with
Provencal authors such as the Nobel Literature
Prize-winner Frederic Mistral, whom he considered
the greatest authority on the Occitan language, and
the other Felibrige poets, as well as his Romanist
colleagues Camille Chabanneau, Emil Levy, Albert
Dauzat, etc. His work was influenced by French
Romanist linguists such as Gaston Paris, Father
Pierre-Jean Rousselot, and Maurice Grammont. He

earned his Doctorat-e`s-Lettres in 1913 from the Sorbonne for his theses Essai de syntaxe des parlers
provencaux modernes and Le developpement du langage observe chez un enfant bilingue. In his later years
he taught as a privat-docent at the University of
Geneva.
In his reviews in the Revue des langues romanes, he
made, as did Grammont, a critical appraisal of the
linguistic geography practiced by Jules Gillieron and
his followers. As denoted by the title (istorique) of
his first thesis, Essai de syntaxe des parlers provencaux modernes, he was a champion of spelling reform
Rorty, Richard (b. 1931) 665
and applied his principles throughout his historical

syntax.
Ronjats posthumously published four-volume
historical syntax of modern Provenc al dialects (the
postmedieval Occitan language), a French government-sponsored publication initiated upon Antoine
Meillets recommendation, appeared under the supervision of Walter von Wartburg and Georges
Millardet.
Ronjat established 19 main features of Occitan: 11
of them are phonetical, 5 morphological, 1 syntactical, and 2 lexical. On this basis, he found 7 linguistic
distinctions between Occitan and Spanish, 8 with
Italian, 12 with Franco-Provenc al, and 16 with
French. He is mainly remembered by Occitan speakers and specialists for his knowledge of all Occitan
dialects. Despite the title (Provenc al spelling), his
study on the Occitan spelling applies to the analysis
of Occitan pronunciation and rules.
In his second thesis on bilingual acquisition,
Le de veloppement du langage observe chez un enfant bilingue, undertaken due to the suggestion of
Grammont, he studied the linguistic development of

his bilingual child, whose mother was German, and
put into practice the linguistic recommendation of
Grammont: one language, one person. This work
is now recognized as a forerunner of contemporary
research on bilingual acquisition (see, for example,
the CHILD Project).
See also: Bilingualism and Second Language Learning;

Occitan; Romance Languages; Syntax of Words.
Bibliography
Ronjat J (1913). Le De veloppement du langage observe
chez un enfant bilingue. Paris: Champion.
Ronjat J (19301941). Grammaire istorique des parlers
provenc aux modernes (4 vols). Montpellier: Socie te des
Langues Romanes.
Ronjat J (1937). LOurtougra`fi Prouvenc alo: pichot tratat a
lusage di Prouvenc au. (segoundo edicioun). Edicioun de
la Mantene`nc o de Prouve`nc o.
Rorty, Richard (b. 1931)

R Chatterjee, Lado International College, Silver
Spring, MD, USA
Richard Rorty was born October 4, 1931, in New

York City. His parents were Trotskyites, and his
early ideas were imbued with a Marxism that he
later outgrew. Rorty is known as a pragmatist, a
social commentator, and a public intellectual. One
writer has chosen the term post-philosophical to
describe his work (Rumana). A precocious intellectual, Rorty enrolled at the University of Chicago when
he was 15 and received his Ph.D. from Yale in 1956.
While teaching at Wellesley and Princeton, Rorty
read deeply in Carnap, Tarski, and Reichenbach. By
this time, many Anglo-American philosophers were
trying to read and understand the later Wittgenstein.
One of Rortys early contributions to philosophy
was an edition of essays by philosophers entitled
The linguistic turn (1967/1992) that runs to some
400 double-columned pages. His introduction,
Metaphysical difficulties of linguistic philosophy,
is often cited. One part of the collection is composed
of essays on the thesis that philosophical questions
are questions of language; another is composed of
problems of ideal-language philosophy, and a third

of ordinary language philosophy. Wittgenstein here is
the 800-pound gorilla who couldnt be represented
because his estate doesnt allow his work to be
excerpted.
Representative of Rortys mature work is his book
Philosophy and the mirror of nature (1979). A critique of epistemology is contained in its earlier parts,
of the notion inherited from the past that it is a
combined development of the ideas of mind, knowledge, and of earlier philosophy itself. This critique
leads in later work, such as Truth and progress
(1998), to the characteristic doctrine of epistemological behaviorism, by which Rorty maintains that
knowledge is constrained only by conversational
constraints. Hence the notion of philosophy as an
edifying conversation. The roots of epistemological
behaviorism Rorty finds in Dewey and in Wittgensteins critique of representation the idea that all
knowledge is found represented in the mind and is
communicable from mind to mind. A further consequence is a denial of any possibility that the nature of
the world constrains us to speak in certain ways. This
view is reminiscent of the Nietzschean critique of
language by which all language is metaphor, or
fiction which is dependent on fictions.
Rorty, Richard (b. 1931) 665
and applied his principles throughout his historical

syntax.
Ronjats posthumously published four-volume
historical syntax of modern Provencal dialects (the
postmedieval Occitan language), a French government-sponsored publication initiated upon Antoine
Meillets recommendation, appeared under the supervision of Walter von Wartburg and Georges
Millardet.
Ronjat established 19 main features of Occitan: 11
of them are phonetical, 5 morphological, 1 syntactical, and 2 lexical. On this basis, he found 7 linguistic
distinctions between Occitan and Spanish, 8 with
Italian, 12 with Franco-Provencal, and 16 with
French. He is mainly remembered by Occitan speakers and specialists for his knowledge of all Occitan
dialects. Despite the title (Provencal spelling), his
study on the Occitan spelling applies to the analysis
of Occitan pronunciation and rules.
In his second thesis on bilingual acquisition,
Le developpement du langage observe chez un enfant bilingue, undertaken due to the suggestion of
Grammont, he studied the linguistic development of

his bilingual child, whose mother was German, and
put into practice the linguistic recommendation of
Grammont: one language, one person. This work
is now recognized as a forerunner of contemporary
research on bilingual acquisition (see, for example,
the CHILD Project).
See also: Bilingualism and Second Language Learning;

Occitan; Romance Languages; Syntax of Words.
Bibliography
Ronjat J (1913). Le Developpement du langage observe
chez un enfant bilingue. Paris: Champion.
Ronjat J (19301941). Grammaire istorique des parlers
provencaux modernes (4 vols). Montpellier: Societe des
Langues Romanes.
Ronjat J (1937). LOurtougra`fi Prouvencalo: pichot tratat a
lusage di Prouvencau. (segoundo edicioun). Edicioun de
la Mantene`nco de Prouve`nco.
Rorty, Richard (b. 1931)

R Chatterjee, Lado International College, Silver
Spring, MD, USA
Richard Rorty was born October 4, 1931, in New

York City. His parents were Trotskyites, and his
early ideas were imbued with a Marxism that he
later outgrew. Rorty is known as a pragmatist, a
social commentator, and a public intellectual. One
writer has chosen the term post-philosophical to
describe his work (Rumana). A precocious intellectual, Rorty enrolled at the University of Chicago when
he was 15 and received his Ph.D. from Yale in 1956.
While teaching at Wellesley and Princeton, Rorty
read deeply in Carnap, Tarski, and Reichenbach. By
this time, many Anglo-American philosophers were
trying to read and understand the later Wittgenstein.
One of Rortys early contributions to philosophy
was an edition of essays by philosophers entitled
The linguistic turn (1967/1992) that runs to some
400 double-columned pages. His introduction,
Metaphysical difficulties of linguistic philosophy,
is often cited. One part of the collection is composed
of essays on the thesis that philosophical questions
are questions of language; another is composed of
problems of ideal-language philosophy, and a third

of ordinary language philosophy. Wittgenstein here is
the 800-pound gorilla who couldnt be represented
because his estate doesnt allow his work to be
excerpted.
Representative of Rortys mature work is his book
Philosophy and the mirror of nature (1979). A critique of epistemology is contained in its earlier parts,
of the notion inherited from the past that it is a
combined development of the ideas of mind, knowledge, and of earlier philosophy itself. This critique
leads in later work, such as Truth and progress
(1998), to the characteristic doctrine of epistemological behaviorism, by which Rorty maintains that
knowledge is constrained only by conversational
constraints. Hence the notion of philosophy as an
edifying conversation. The roots of epistemological
behaviorism Rorty finds in Dewey and in Wittgensteins critique of representation the idea that all
knowledge is found represented in the mind and is
communicable from mind to mind. A further consequence is a denial of any possibility that the nature of
the world constrains us to speak in certain ways. This
view is reminiscent of the Nietzschean critique of
language by which all language is metaphor, or
fiction which is dependent on fictions.
666 Rorty, Richard (b. 1931)
Apart from his academic activity, Rorty is engaged

as a public intellectual and activist. This engagement
is launched on an anti-essentialist basis. The intellectual or philosopher combines a realization of the
contingency of his or her own vocabulary with the
goals of reducing suffering and cruelty. The vehicle of
activism is not argument, which has been rendered
void by the pragmatist critique of reason, but a redescription or realignment of vocabularies in a
continuing conversation. Here Rorty distinguishes
himself from any radicalism that seeks to detect
hitherto unknown reasons for social ills as a warrant
for removing them. It is significant that most of Rortys work is achieved in conversation with linguistic
and analytic philosophers such as Wittgenstein,
Quine, Sellars, Davidson, and Putnam.
See also: Carnap, Rudolf (18911970); Davidson, Donald
(19172003); Putnam, Hilary (b. 1926); Quine, Willard van
Orman (19082000); Reichenbach, Hans (18911953); Romanian; Tarski, Alfred (19021983).
Bibliography
Brandom R (ed.) (2000). Richard Rorty: the philosopher
meets his critics. Oxford/Cambridge, MA: Blackwell.
Rorty R (ed.) (1967/1992). The linguistic turn (with a
new introduction). Chicago: The University of Chicago
Press.
Rorty R (1979). Philosophy and the mirror of nature. Princeton, NJ: Princeton University Press.
Rorty R (1991a). Objectivity, relativism and truth: philosophical papers 1. Cambridge: Cambridge University
Press.
Rorty R (1991b). Essays on Heidegger and others: philosophical papers 2. Cambridge: Cambridge University
Press.
Rorty R (1998). Truth and progress: philosophical papers 3.
Rosetti, Alexandru (18951990)

G Price
This article is reproduced from the previous edition, volume 7,
p. 3612, 1994, Elsevier Ltd.
Alexandru Rosetti was born in Bucharest on October

23, 1895. After graduating in 1920 at the University
of Bucharest he spent eight years in Paris where he
studied under Meillet, Rousselot, and Gillie ron (see
Meillit, Antoine (Paul Jules) (18661936); Rousselot,
Pierre Jean, Abbe (18461924); Gillieron, Jules
(18541926)), among others and obtained his doctorate in 1926 for a thesis on Romanian historical phonetics. He returned to Romania in 1928 to take up a
post at the University of Bucharest and in 1932 was
promoted to the rank of professor. In 1938 he succeeded his former teacher, Ovid Densusianu, in the
chair of the Romanian language which he was to
occupy until his retirement in 1965. He maintained
throughout his long life his preoccupation with phonetics, both historical and theoretical, and directed
(19631969) the Center for Phonetic and Dialectal
Research in Bucharest. His authoritative history of
the Romanian language up to the 16th century,
which was first published in six parts (19381966),
was revised in one volume, taking it up to the 17th
century, in 1968 (3rd edn., 1986). He also made
important contributions to general linguistic theory,

in particular with his book (1943) tackling the problem of defining the word as a linguistic entity. He
founded and edited the periodical Bulletin linguistique (19331948) and later, when existing journals
were discontinued and replaced by others after the
communist regime took full control of the country in
1948, edited at various times the periodicals Studii s i
cercetari lingvistice; Fonetica si dialectologie; Revue
roumaine de linguistique; and Cahiers de linguistique
theorique et appliquee. In the 1930s he was closely
associated with King Carol II and was editorial director (19331945) of the Royal Foundation for Literature and Art. He succeeded, however, in coexisting
with the post-1948 communist regime (he was, indeed, one of the foundation members of the Academy
of the Romanian Peoples Republic which replaced
the Romanian Academy in that year), but (unlike
some others) without compromising his intellectual
probity. He outlived the regime by two months and
one of his last acts, after the overthrow of Nicolae
Ceaus escu in December 1989, was to call for the full
restoration of the Institute of Phonetics, which, like
various other institutions, had been downgraded.
This was done and, after his death in Bucharest on
February 27, 1990, the Institute was renamed after
him. (Bibliographies of his writings up to 1965, 1975,
and 1980, respectively, may be found in Rosetti,
1965, 1975, 1980.)
666 Rorty, Richard (b. 1931)
Apart from his academic activity, Rorty is engaged

as a public intellectual and activist. This engagement
is launched on an anti-essentialist basis. The intellectual or philosopher combines a realization of the
contingency of his or her own vocabulary with the
goals of reducing suffering and cruelty. The vehicle of
activism is not argument, which has been rendered
void by the pragmatist critique of reason, but a redescription or realignment of vocabularies in a
continuing conversation. Here Rorty distinguishes
himself from any radicalism that seeks to detect
hitherto unknown reasons for social ills as a warrant
for removing them. It is significant that most of Rortys work is achieved in conversation with linguistic
and analytic philosophers such as Wittgenstein,
Quine, Sellars, Davidson, and Putnam.
See also: Carnap, Rudolf (18911970); Davidson, Donald
(19172003); Putnam, Hilary (b. 1926); Quine, Willard van
Orman (19082000); Reichenbach, Hans (18911953); Romanian; Tarski, Alfred (19021983).
Bibliography
Brandom R (ed.) (2000). Richard Rorty: the philosopher
meets his critics. Oxford/Cambridge, MA: Blackwell.
Rorty R (ed.) (1967/1992). The linguistic turn (with a
new introduction). Chicago: The University of Chicago
Press.
Rorty R (1979). Philosophy and the mirror of nature. Princeton, NJ: Princeton University Press.
Rorty R (1991a). Objectivity, relativism and truth: philosophical papers 1. Cambridge: Cambridge University
Press.
Rorty R (1991b). Essays on Heidegger and others: philosophical papers 2. Cambridge: Cambridge University
Press.
Rorty R (1998). Truth and progress: philosophical papers 3.
Rosetti, Alexandru (18951990)

G Price
p. 3612, 1994, Elsevier Ltd.
Alexandru Rosetti was born in Bucharest on October

23, 1895. After graduating in 1920 at the University
of Bucharest he spent eight years in Paris where he
studied under Meillet, Rousselot, and Gillieron (see
Meillit, Antoine (Paul Jules) (18661936); Rousselot,
Pierre Jean, Abbe (18461924); Gillieron, Jules
(18541926)), among others and obtained his doctorate in 1926 for a thesis on Romanian historical phonetics. He returned to Romania in 1928 to take up a
post at the University of Bucharest and in 1932 was
promoted to the rank of professor. In 1938 he succeeded his former teacher, Ovid Densusianu, in the
chair of the Romanian language which he was to
occupy until his retirement in 1965. He maintained
throughout his long life his preoccupation with phonetics, both historical and theoretical, and directed
(19631969) the Center for Phonetic and Dialectal
Research in Bucharest. His authoritative history of
the Romanian language up to the 16th century,
which was first published in six parts (19381966),
was revised in one volume, taking it up to the 17th
century, in 1968 (3rd edn., 1986). He also made
important contributions to general linguistic theory,

in particular with his book (1943) tackling the problem of defining the word as a linguistic entity. He
founded and edited the periodical Bulletin linguistique (19331948) and later, when existing journals
were discontinued and replaced by others after the
communist regime took full control of the country in
1948, edited at various times the periodicals Studii si
cercetari lingvistice; Fonetica si dialectologie; Revue
roumaine de linguistique; and Cahiers de linguistique
theorique et appliquee. In the 1930s he was closely
associated with King Carol II and was editorial director (19331945) of the Royal Foundation for Literature and Art. He succeeded, however, in coexisting
with the post-1948 communist regime (he was, indeed, one of the foundation members of the Academy
of the Romanian Peoples Republic which replaced
the Romanian Academy in that year), but (unlike
some others) without compromising his intellectual
probity. He outlived the regime by two months and
one of his last acts, after the overthrow of Nicolae
Ceausescu in December 1989, was to call for the full
restoration of the Institute of Phonetics, which, like
various other institutions, had been downgraded.
This was done and, after his death in Bucharest on
February 27, 1990, the Institute was renamed after
him. (Bibliographies of his writings up to 1965, 1975,
and 1980, respectively, may be found in Rosetti,
1965, 1975, 1980.)
Rousseau, Jean-Jacques (17121778) 667

See also: Gillieron, Jules (18541926); Meillit, Antoine
(Paul Jules) (18661936); Rousselot, Pierre Jean, Abbe

(18461924).
Bibliography
Rosetti A (1926). Recherches sur la phone tique du roumain
au XVIe sie`cle. Paris: Champion.
Rosetti A (1947). Le Mot: Esquisse dune the orie ge ne rale.
Copenhagen: Munksgaard.
Rosetti A, Cazacu B & Onu L (1961). Istoria limbii romne

literare: De la origini pna la nceputul secolului al XIXlea. Bucharest: Editura s tiintifica.
Rosetti A (1965). Omagiu lu: Alexandru Rosetti la 70 de
ani. Bucharest: Editura Academiei.
Rosetti A (1968). Istoria limbii roma ne: De la origini pna
n secolul al XVII-lea. Bucharest: Editura stiintifica.
Rosetti A (1975). Bibliography. Revue Roumaine de
Linguistique 20, 445449.
Rousseau, Jean-Jacques (17121778)

D Droixhe
pp. 36123613, 1994, Elsevier Ltd.
Rousseaus Essai sur lorigine des langues mixes in a

paradoxical way favorite topics of the 18th century
and passionate but also hesitating intuitions about
the politics of language.
The Discours sur lorigine de line galite (1755) stresses the embarrassments of the question of the origin of
language (see Origin of Language Debate), in a manner reminiscent of Frain du Tremblay (Traite des langues, 1703). Two types of conclusions were deduced:
(a) this question is unsolvable and the efforts of such a
great mind are typically vain (Kant; see Kant, Immanuel (17241804); Nietzsche, etc); (b) there is an absolute solidarity between language and society, from
which some authors will infer a sort of linguistic revelation related to a fixed political order.
According to Rousseau, the Essai sur lorigine des
langues was a fragment of the Discours sur line galite , cut out for being too long and irrelevant. In so
far as the chapters of the Essai dealing with the politics of language offer ideas differing from the Discours, it has been supposed that they give an earlier,
immature state of Rousseaus thinking. The contrary
was also argued: the passage cut out would in fact be
dealing with the polemic which opposed to Rameaus
theory of harmony Rousseaus ideas about the melodic musicality of Italian, supposedly lost by French
and more suitable to opera.
Rousseaus complexities or contradictions about
language can be traced back to the picture of the
origins given in the Discours. Speech arises amid
singing and dancing true children of love and leisure. The best singer or dancer, the one who was the
most beautiful, strong, skilful, or eloquent became

the most esteemed; and it was the first step towards
inequality, and towards vice at the same time. An
idealistic, sentimental conception of language is combined at least in a dialectic tension with a critical
and materialistic approach associating language with
the rise of social damage. Words appear in an order
radically heterogeneous to the natural one, endowed
with an irreducible autonomy and originality. Modern criticism has stressed the fact that the birth of a
real language is described as it took place in historical
times, and then seems to be separated from the very
principle of origin (Starobinski, 1990).
The link between speech, passion, and singing,
based on a climatology opposing the northern languages to the south, is partly borrowed from Condillacs genesis of human arts and polemically directed
against his giving a prominent part to material needs
(see Condillac, Etienne Bonnot de (17141780)). Any
paradisiac beginnings of man are revoked in the Essai,
as well as a linear development of language, which
seems to have more complex origins. The ages of the
hunters (bloodthirsty barbarians) and of the shepherds are characterized by social division and they are
supposed to prepare the rise of a real civil language
appearing with agriculture. On one side, the experiment of mutual help which required some communication seems to participate at each level in the
transcending of animal selfishness towards the accomplishment of mans social nature. On the other, the
Essai puts a critical emphasis on the depravation of
urbane language, which became a cold and courteous
instrument of oppression.
The first signs were primitive metaphors (see
Metaphor: Philosophical Theories). This preromantic topic, perhaps stemming from Vico, found its full
rationality in sensualist philosophy, which sees words
as dynamic and culturally signifying images of the
world, working on reality to give it some sense, just as
Rousseau, Jean-Jacques (17121778) 667

See also: Gillieron, Jules (18541926); Meillit, Antoine
(Paul Jules) (18661936); Rousselot, Pierre Jean, Abbe

(18461924).
Bibliography
Rosetti A (1926). Recherches sur la phonetique du roumain
au XVIe sie`cle. Paris: Champion.
Rosetti A (1947). Le Mot: Esquisse dune theorie generale.
Copenhagen: Munksgaard.
Rosetti A, Cazacu B & Onu L (1961). Istoria limbii romne

literare: De la origini pna la nceputul secolului al XIXlea. Bucharest: Editura stiintifica.
Rosetti A (1965). Omagiu lu: Alexandru Rosetti la 70 de
ani. Bucharest: Editura Academiei.
Rosetti A (1968). Istoria limbii romane: De la origini pna
n secolul al XVII-lea. Bucharest: Editura stiintifica.
Rousseau, Jean-Jacques (17121778)

D Droixhe
pp. 36123613, 1994, Elsevier Ltd.
Rousseaus Essai sur lorigine des langues mixes in a

paradoxical way favorite topics of the 18th century
and passionate but also hesitating intuitions about
the politics of language.
The Discours sur lorigine de linegalite (1755) stresses the embarrassments of the question of the origin of
language (see Origin of Language Debate), in a manner reminiscent of Frain du Tremblay (Traite des langues, 1703). Two types of conclusions were deduced:
(a) this question is unsolvable and the efforts of such a
great mind are typically vain (Kant; see Kant, Immanuel (17241804); Nietzsche, etc); (b) there is an absolute solidarity between language and society, from
which some authors will infer a sort of linguistic revelation related to a fixed political order.
According to Rousseau, the Essai sur lorigine des
langues was a fragment of the Discours sur linegalite, cut out for being too long and irrelevant. In so
far as the chapters of the Essai dealing with the politics of language offer ideas differing from the Discours, it has been supposed that they give an earlier,
immature state of Rousseaus thinking. The contrary
was also argued: the passage cut out would in fact be
dealing with the polemic which opposed to Rameaus
theory of harmony Rousseaus ideas about the melodic musicality of Italian, supposedly lost by French
and more suitable to opera.
Rousseaus complexities or contradictions about
language can be traced back to the picture of the
origins given in the Discours. Speech arises amid
singing and dancing true children of love and leisure. The best singer or dancer, the one who was the
most beautiful, strong, skilful, or eloquent became

the most esteemed; and it was the first step towards
inequality, and towards vice at the same time. An
idealistic, sentimental conception of language is combined at least in a dialectic tension with a critical
and materialistic approach associating language with
the rise of social damage. Words appear in an order
radically heterogeneous to the natural one, endowed
with an irreducible autonomy and originality. Modern criticism has stressed the fact that the birth of a
real language is described as it took place in historical
times, and then seems to be separated from the very
principle of origin (Starobinski, 1990).
The link between speech, passion, and singing,
based on a climatology opposing the northern languages to the south, is partly borrowed from Condillacs genesis of human arts and polemically directed
against his giving a prominent part to material needs
(see Condillac, Etienne Bonnot de (17141780)). Any
paradisiac beginnings of man are revoked in the Essai,
as well as a linear development of language, which
seems to have more complex origins. The ages of the
hunters (bloodthirsty barbarians) and of the shepherds are characterized by social division and they are
supposed to prepare the rise of a real civil language
appearing with agriculture. On one side, the experiment of mutual help which required some communication seems to participate at each level in the
transcending of animal selfishness towards the accomplishment of mans social nature. On the other, the
Essai puts a critical emphasis on the depravation of
urbane language, which became a cold and courteous
instrument of oppression.
The first signs were primitive metaphors (see
Metaphor: Philosophical Theories). This preromantic topic, perhaps stemming from Vico, found its full
rationality in sensualist philosophy, which sees words
as dynamic and culturally signifying images of the
world, working on reality to give it some sense, just as
668 Rousseau, Jean-Jacques (17121778)
metaphors transform things and designations by adding new meanings. This paved the way for Humboldtian relativism.
The Discours de line galite had the provocative
sharpness of a reference pattern opposing ideal to
reality. The Essai produced a more complex picture
which introduced violence and economical rationality of history into the origins of language, lost some of
its critical power and left the genetic problem unsolved. It is understandable that Rousseau did not
publish his scribble (edited posthumously in 1781),
while he was personally experimenting with the failures and ambiguous nature of language, always oscillating between individual and community (Emile),
and the regeneration of a style challenging the heat

of natural communication.
See also: Condillac, Etienne Bonnot de (17141780); Kant,
Immanuel (17241804); Metaphor: Philosophical Theories; Origin of Language Debate.
Bibliography
Porset C (1976). Linque tante e trangete de LEssai sur
lorigine des langues: Rousseau et ses interpre`tes. Studies
on Voltaire and the eighteenth century.
Starobinski J (1990). Pre sentation. In Rousseau J J (ed.)
Essai sur lorigine des langues. Paris: Gallimard.
Rousselot, Pierre Jean, Abbe (18461924)

C Puech, Universite Sorbonne Nouvelle Paris III,
Paris, France
Rousselot was born on October 10, 1846 in SaintClaude (Charente region) to a rural family. He
received a classical education at the seminary in
Richemont, then at Angoule me, and was ordained
priest in 1970. He first studied Latin and Greek phonetics before turning to dialectology. Planning to
complete a doctoral thesis under the supervision of
Gaston Paris, he conducted surveys in the Limousin
region in order to determine the borders of langue
doc and langue dol (names of the two principal
groups of medieval French dialects, named after
their respective words for yes), but suddenly fell
seriously ill. While he was being looked after, mainly
by his mother, in his village, he wrote down the
phonetic differences in the provincial dialect used by
the various members of his family. His observations,
combined with the laboratory methods of articulatory recording and acoustic analysis designed by physiologists Marey and Rosapelly, enabled him to
defend his thesis in 1891: Les modifications phone tiques du langage e tudie es dans le patois dune famille
de Cellefrouin (The phonetic modifications of language, studied in the provincial dialect spoken by a
family in Cellefrouin). This work is a diachronic
analysis of a variety of ol as spoken by three generations of the same family, from the point of view of
physiological phonetics and acoustics as far as those
perspectives are relevant to the description of language change. He advocates distance from written
language forms and urges dialectologists to study
the living speech from oral sources, to take into

account such features as thought over, instinctive,
ideal and social characteristics (age, place of birth,
language habits, etc.) In the first issue of the Revue
des patois gallo-romans (18871892), which he coedited with Jules Gillie ron, he published an important
Introduction to the study of provincial dialects. He
developed his instrumental, physiological, and physical analysis of sounds in a treaty that was considerably enriched between the first manuscript, which
was awarded the Volney Prize in 1896, and the last
edition (Principles of experimental phonetics). After
he taught at the Institut Catholique de Paris in 1882,
Michel Bre al created for him an experimental phonetics laboratory in connection with his Chair of
Comparative Literature at the Colle`ge de France. It
was only in 1922 that Rousselot obtained a teaching
position, and then a chair in experimental phonetics
that he was to keep only one year.
Rousselot put phonetics at the service of medicine.
Together with Dr Marcel Natier he founded the Institut dOrthophonie et de Laryngologie and the journal
La Parole (18991904). He did applied research on
various subjects (nose, ears, throat, larynx, stuttering,
etc.). He also played a founding part in the fields of
didactics and foreign language teaching. He applied
his subject to orthoepy in courses given at the Alliance
Franc aise in Paris (Centre de correction des accents
e trangers, Centre for Correction of Foreign Accents,
founded in 1897) as well as in Germany, and through
the publishing of the Pre cis de prononciation franc aise (Handbook of French pronunciation), which
he wrote with his nephew Fauste Laclotte. He also
published a series of articles on teaching through
eyesight, after a (phono-technical) method based on
668 Rousseau, Jean-Jacques (17121778)
metaphors transform things and designations by adding new meanings. This paved the way for Humboldtian relativism.
The Discours de linegalite had the provocative
sharpness of a reference pattern opposing ideal to
reality. The Essai produced a more complex picture
which introduced violence and economical rationality of history into the origins of language, lost some of
its critical power and left the genetic problem unsolved. It is understandable that Rousseau did not
publish his scribble (edited posthumously in 1781),
while he was personally experimenting with the failures and ambiguous nature of language, always oscillating between individual and community (Emile),
and the regeneration of a style challenging the heat

of natural communication.
See also: Condillac, Etienne Bonnot de (17141780); Kant,
Immanuel (17241804); Metaphor: Philosophical Theories; Origin of Language Debate.
Bibliography
Porset C (1976). Linquetante etrangete de LEssai sur
lorigine des langues: Rousseau et ses interpre`tes. Studies
on Voltaire and the eighteenth century.
Starobinski J (1990). Presentation. In Rousseau J J (ed.)
Essai sur lorigine des langues. Paris: Gallimard.
Rousselot, Pierre Jean, Abbe (18461924)

C Puech, Universite Sorbonne Nouvelle Paris III,
Paris, France
Rousselot was born on October 10, 1846 in SaintClaude (Charente region) to a rural family. He
received a classical education at the seminary in
Richemont, then at Angouleme, and was ordained
priest in 1970. He first studied Latin and Greek phonetics before turning to dialectology. Planning to
complete a doctoral thesis under the supervision of
Gaston Paris, he conducted surveys in the Limousin
region in order to determine the borders of langue
doc and langue dol (names of the two principal
groups of medieval French dialects, named after
their respective words for yes), but suddenly fell
seriously ill. While he was being looked after, mainly
by his mother, in his village, he wrote down the
phonetic differences in the provincial dialect used by
the various members of his family. His observations,
combined with the laboratory methods of articulatory recording and acoustic analysis designed by physiologists Marey and Rosapelly, enabled him to
defend his thesis in 1891: Les modifications phonetiques du langage etudiees dans le patois dune famille
de Cellefrouin (The phonetic modifications of language, studied in the provincial dialect spoken by a
family in Cellefrouin). This work is a diachronic
analysis of a variety of ol as spoken by three generations of the same family, from the point of view of
physiological phonetics and acoustics as far as those
perspectives are relevant to the description of language change. He advocates distance from written
language forms and urges dialectologists to study
the living speech from oral sources, to take into

account such features as thought over, instinctive,
ideal and social characteristics (age, place of birth,
language habits, etc.) In the first issue of the Revue
des patois gallo-romans (18871892), which he coedited with Jules Gillieron, he published an important
Introduction to the study of provincial dialects. He
developed his instrumental, physiological, and physical analysis of sounds in a treaty that was considerably enriched between the first manuscript, which
was awarded the Volney Prize in 1896, and the last
edition (Principles of experimental phonetics). After
he taught at the Institut Catholique de Paris in 1882,
Michel Breal created for him an experimental phonetics laboratory in connection with his Chair of
Comparative Literature at the Colle`ge de France. It
was only in 1922 that Rousselot obtained a teaching
position, and then a chair in experimental phonetics
that he was to keep only one year.
Rousselot put phonetics at the service of medicine.
Together with Dr Marcel Natier he founded the Institut dOrthophonie et de Laryngologie and the journal
La Parole (18991904). He did applied research on
various subjects (nose, ears, throat, larynx, stuttering,
etc.). He also played a founding part in the fields of
didactics and foreign language teaching. He applied
his subject to orthoepy in courses given at the Alliance
Francaise in Paris (Centre de correction des accents
etrangers, Centre for Correction of Foreign Accents,
founded in 1897) as well as in Germany, and through
the publishing of the Precis de prononciation francaise (Handbook of French pronunciation), which
he wrote with his nephew Fauste Laclotte. He also
published a series of articles on teaching through
eyesight, after a (phono-technical) method based on
Rousselot, Pierre Jean, Abbe (18461924) 669
graphical lines recording organic activity (La Parole,

1901, 1902).
Although he contributed to bringing phonetics, dialectology, physics, and physiology closer together,
Rousselot suffered from a lack of recognition while
he was alive. He was not able to impose his methods
either in the institutions for which he had conducted
his research (music and drama academies, language
institutes, clinics) or in higher education. When he
died (in Paris, on December 16, 1924), his chair at
the Colle`ge de France was turned into a chair of
histo-physiology, and most of the phonetic equipment
was recovered by the Institut Catholique. His system
of phonetic transcription a direct competitor of the
IPA is still used, though by French specialists in
Romance languages.
See also: Breal, Michel Jules Alfred (18321915); Gillieron,
Jules (18541926); Volney, Constantin-Francois Chasseboeuf, Comte de (17571820).
Bibliography
Asher R E & Henderson Euge nie J A (1981). Towards a
history of phonetics. Edinburgh: Edinburgh University
Press.
Auroux S (1979). La cate gorie duparler et la linguistique.
In Romantisme. Revue du dix-neuvie`me sie`cle, 2526.
Paris: C. D. U. et SEDES. 157178.
Baggioni D (1996). Rousselot, Abbe Jean-Pierre. In
Stammerjohann H et al. (eds.) Lexicon grammaticorum.
Whos who in the history of world linguistics. Tu bingen:
Niemeyer. 806.
Bergounioux G (1994). Aux origines de la linguistique
franc aise. Textes choisis et pre sente s. Paris: Pocket, coll.
Agora Les Classiques.
Bergounioux G (1996). Phone tique et dialectologie au XIXe
sie`cle. In Lanalisi linguistica e letteraria, anne e IV, 1.
Milan: Vita e Pensiero Publications de lUniversita`
Cattolica del Sacro Cuore. 2746.
Boe L-J (2000). Notice 5404. Rousselot, Abbe Jean-Pierre,
Principes de phone tique expe rimentale. In Colombat B
& assiste de Lazcano E (eds.) Corpus repre sentatif des
grammaires et des traditions linguistiques, tome 2, Histoire Episte mologie Langage Hors se rie 3. Paris: SHESL.
509512.
Bronstein A J, Raphael L J & Stevens C J (1977). A biographical dictionary of the phonetic sciences. New York:
The Press of Lehman College.
Brunot F (1925). La phone tique expe rimentale. Le Monde
Illustre , 31 janvier 1925, 7273.
Carton F (1995). La phone tique expe rimentale, la phonologie, les archives sonores. In Antoine G & Martin R
(eds.) Histoire de la langue franc aise 19141945. Paris:
CNRS. 873894.
Gallazzi E (1995). Phone tique/Universite /Enseignement
des langues a` la fin du XIXe sie`cle. Histoire Episte mologie Language 17I, 95114.
Gallazzi E (1996). LA.P.I. a` ses de buts: alphabet universel

et variations. In Lanalisi linguistica e letteraria, anne e
IV, 1. Milan: Vita e Pensiero Publications de
lUniversita` Cattolica del Sacro Cuore.
Gallazzi E (2000). Lassociation phone tique internationale.
In Auroux S (ed.) Histoire des ide es linguistiques T. 3
Lhe ge monie du comparatisme. Lie`ge: Pierre Mardaga.
499516.
Gallazzi E (2001). Physiologie de la parole et phone tique
applique e au 19e et au de but du 20e sie`cle. In Auroux S,
Koerner E F K, Niederehe H-J & Versteegh K (eds.)
History of the language sciences: an international handbook on the evolution of the study of language from the
beginnings to the present Vol. II. Berlin, New York:
Walter de Gruyter. 14851498.
Koerner E F K (1994). Introduction: towards a history
of phonetics. In Panconcelli-Calzia G (ed.) Geschichtszahlen der Phonetik, Quellenatlas der Phonetik (new
edition with an English introduction). Amsterdam and
Philadelphia: J. Benjamins (SiHoL, 16). ixxxxviii.
Malmberg B (1991). Histoire de la linguistique de Sumer a`
Saussure. Paris: PUF.
Millet Abbe (1925a). Pre cis dexpe rimentation phone tique.
La physiologie des articulations I. Enregistrement. II.
Interpre tation (avec 52 figures dans le texte) pre ce de
dune Introduction sur luvre de labbe Rousselot.
Paris, Toulouse: Henri Didier, Edouard Privat. 1133.
[La date de dition est 1926 sur la couverture, 1925 sur
la page de titre.]
Millet Abbe (1925b). Luvre de labbe Rousselot. La
science expe rimentale du langage. In Pre cis dexpe rimentation phone tique. La physiologie des articulations
I. Enregistrement. II. Interpre tation (avec 52 figures
dans le texte) pre ce de dune Introduction sur luvre de
labbe Rousselot. Paris, Toulouse: Henri Didier, Edouard
Privat. 1133.
Millet Abbe (1925c). Liste des publications de labbe
Rousselot. In Pre cis dexpe rimentation phone tique. La
physiologie des articulations I. Enregistrement. II. Interpre tation (avec 52 figures dans le texte) pre ce de dune Introduction sur luvre de labbe Rousselot. Paris, Toulouse:
Henri Didier, Edouard Privat. IIII (en fin douvrage).
Moses E R (1964). Phonetics. History and interpretation.
Englewood Cliffs, NJ: Prentice Hall.
Pop S (1950). La dialectologie. Aperc u historique et me thodes denque tes linguistiques. 1e`re partie. Dialectologie
romane.2e partie. Dialectologie non romane. Louvain:
Impr. Duculot, Gembloux.
Pop S (ed.) (1956). Instituts de Phone tique et Archives
phonographiques. Louvain: Publications de la Commission denque te linguistique.
Rousselot Abbe (1887). Introduction a` le tude des patois.
In Revue des patois gallo-romans, T. 1. Paris, Neucha tel:
Champion, Attinger. 122.
Rousselot Abbe (1891). Les modifications phone tiques du
langage e tudie es dans le patois dune famille de Cellefrouin (Charente), The`se pre sente e a` la faculte des lettres
de Paris. Paris: H. Welter.
Rousselot Abbe (1897, 1901, 1908). Principes de phone tique expe rimentale. Paris: H. Welter.
670 Rousselot, Pierre Jean, Abbe (18461924)

Rousselot Abbe , & Laclotte F (1902). Pre cis de prononciation franc aise. Paris: H. Welter.
Rousselot Abbe (19241925). Principes de phone tique
expe rimentale, e dition augmente e. T. 1: Paris: Didier,
T. 2: Paris, Toulouse: Didier, Privat.
Souza R de (1925). Labbe Rousselot et la phone tique
expe rimentale. Une grande uvre scientifique en pe ril.
In Le Correspondant: religion, philosophie, politique,

histoire, sciences, e conomie sociale, voyage, litte rature,
beaux-arts, bimensuel, tome 300e, 5e livraison (10
Septembre 1925), 97e anne e. Paris: Bureaux du Correspondant. 666691.
Tagliavini C (1963). Panorama di storia della linguistica.
Bologne: Riccardo Pa`tron.
Rozwadowski, Jan Micha (18671935)

Z Wasik, Adam Mickiewicz University, Poznan, Poland
(
Jan Micha [Jordan] Rozwadowski was born in

Czarna, near Tarno w (at that time Austro-Hungarian
Galitia), on December 7, 1867. He died on March 14,
1935, on the train, on his return from Warsaw following a lecture in Cracow. Rozwadowski had held
a Professorship in Comparative Grammar at the
Jagiellonian University, and his first philological training began in Cracow with Latin and Sanskrit. In 1890,
he passed an exam entitling him to teach in an
elementary school, and a year later he received his
Ph.D. To study linguistics, he applied for a scholarship in Germany in 1891, and again in 1894. His academic experience was enriched through the lectures,
inter alia, of the Slavicist A. Leskien, the Germanist
H. Hirt, and the phonetician E. Sievers, in Leipzig.
Rozwadowski primarily took part in seminars conducted by K. Brugmann, under whose supervision
he wrote several works, including his Habilitationsschrift (reviewed later on by J. N. Baudouin de
Courtenay), defended as Questiones grammaticae et
etymologicae, in conjunction with a lecture on the
causes of change of word meaning in the summer of
1897. As an Indo-Europeanist and Slavicist, he made
his name known through numerous works relating
to almost all domains of the sciences of language:
descriptive phonetics, the history of sound changes
and lexical borrowings, dialectology, word formation
and inflection, etymology and semasiology, syntax and
stylistics, and toponomastics. Some of his works dealt
with historical documents of the Polish language,
Polish onomastics and phonetics, and some with
the localization of the homeland of Slavs and IndoEuropeans. He also discovered some regular rules in
the language in general.
First, Rozwadowski formulated in opposition to
W. Wundt, who claimed that extra-linguistic objects
are usually named after one unique feature the law
of bipartition stating that the base of every language
product consists of two parts: a distinguishing component pertaining to its semantic aspect; and an identifying component pertaining to its formal aspect. Second,
he formulated the law of quantitative development of
a language, stating that the phonetic and semantic
difference between pairs of linguistic products must
present a constant quantity: when it diminishes under
this measure, one of the products, as independent,
disappears; when it augments above this quantity, a
new independent product comes into existence. Finally,
Rozwadowski formulated the law of disautomatization, stating that the principal cause of evolution in
linguistic creativity, just as in cultural creativity, is a
constant need for the refreshment of an emotional
factor being used up through automatization.
Widely recognized, Rozwadowski assumed a leading role in the organization of Polish institutions. The
Commission of Language of the Academy of Learning
was created in 1874; later, in 1898, he became its
second secretary. He then held the post of President
in the years between 1908 and 1935. In the meantime, he also fulfilled the duties of vice-president,
beginning in 1920, and afterwards president of the
Polish Academy of Learning from 1925 to 1929.
Since December 10, 1911, his name had been mentioned among the Corresponding Members of the
Division of Russian Language and Philology in The
Imperial St. Petersburg Academy of Sciences (between
1917 and 1925); and after 1991, the Russian Academy of Sciences. On his initiative, on May 31, 1925,
the general assembly of Polish linguists gathered at
Lvov and decided to launch a society, which was to
meet annually and publish a journal devoted to general linguistics. Subsequently, he became elected as the
first and the fourth president of the Polish Linguistic
Society (19251928 and 19301931). Being at the
peak of his scientific accomplishment Rozwadowski
delivered a lecture before the Socie te de Linguistique
de Paris in 1925. He had regarded linguistics as in the
highest degree predestined to formulate generalizations about the mechanisms of human language and
mind; subsequently, he proposed, as the best way to
670 Rousselot, Pierre Jean, Abbe (18461924)

Rousselot Abbe, & Laclotte F (1902). Precis de prononciation francaise. Paris: H. Welter.
Rousselot Abbe (19241925). Principes de phonetique
experimentale, edition augmentee. T. 1: Paris: Didier,
T. 2: Paris, Toulouse: Didier, Privat.
Souza R de (1925). Labbe Rousselot et la phonetique
experimentale. Une grande uvre scientifique en peril.
In Le Correspondant: religion, philosophie, politique,

histoire, sciences, economie sociale, voyage, litterature,
beaux-arts, bimensuel, tome 300e, 5e livraison (10
Septembre 1925), 97e annee. Paris: Bureaux du Correspondant. 666691.
Tagliavini C (1963). Panorama di storia della linguistica.
Bologne: Riccardo Pa`tron.
Rozwadowski, Jan Micha (18671935)

Z Wasik, Adam Mickiewicz University, Poznan, Poland
(
Jan Micha [Jordan] Rozwadowski was born in

Czarna, near Tarnow (at that time Austro-Hungarian
Galitia), on December 7, 1867. He died on March 14,
1935, on the train, on his return from Warsaw following a lecture in Cracow. Rozwadowski had held
a Professorship in Comparative Grammar at the
Jagiellonian University, and his first philological training began in Cracow with Latin and Sanskrit. In 1890,
he passed an exam entitling him to teach in an
elementary school, and a year later he received his
Ph.D. To study linguistics, he applied for a scholarship in Germany in 1891, and again in 1894. His academic experience was enriched through the lectures,
inter alia, of the Slavicist A. Leskien, the Germanist
H. Hirt, and the phonetician E. Sievers, in Leipzig.
Rozwadowski primarily took part in seminars conducted by K. Brugmann, under whose supervision
he wrote several works, including his Habilitationsschrift (reviewed later on by J. N. Baudouin de
Courtenay), defended as Questiones grammaticae et
etymologicae, in conjunction with a lecture on the
causes of change of word meaning in the summer of
1897. As an Indo-Europeanist and Slavicist, he made
his name known through numerous works relating
to almost all domains of the sciences of language:
descriptive phonetics, the history of sound changes
and lexical borrowings, dialectology, word formation
and inflection, etymology and semasiology, syntax and
stylistics, and toponomastics. Some of his works dealt
with historical documents of the Polish language,
Polish onomastics and phonetics, and some with
the localization of the homeland of Slavs and IndoEuropeans. He also discovered some regular rules in
the language in general.
First, Rozwadowski formulated in opposition to
W. Wundt, who claimed that extra-linguistic objects
are usually named after one unique feature the law
of bipartition stating that the base of every language
product consists of two parts: a distinguishing component pertaining to its semantic aspect; and an identifying component pertaining to its formal aspect. Second,
he formulated the law of quantitative development of
a language, stating that the phonetic and semantic
difference between pairs of linguistic products must
present a constant quantity: when it diminishes under
this measure, one of the products, as independent,
disappears; when it augments above this quantity, a
new independent product comes into existence. Finally,
Rozwadowski formulated the law of disautomatization, stating that the principal cause of evolution in
linguistic creativity, just as in cultural creativity, is a
constant need for the refreshment of an emotional
factor being used up through automatization.
Widely recognized, Rozwadowski assumed a leading role in the organization of Polish institutions. The
Commission of Language of the Academy of Learning
was created in 1874; later, in 1898, he became its
second secretary. He then held the post of President
in the years between 1908 and 1935. In the meantime, he also fulfilled the duties of vice-president,
beginning in 1920, and afterwards president of the
Polish Academy of Learning from 1925 to 1929.
Since December 10, 1911, his name had been mentioned among the Corresponding Members of the
Division of Russian Language and Philology in The
Imperial St. Petersburg Academy of Sciences (between
1917 and 1925); and after 1991, the Russian Academy of Sciences. On his initiative, on May 31, 1925,
the general assembly of Polish linguists gathered at
Lvov and decided to launch a society, which was to
meet annually and publish a journal devoted to general linguistics. Subsequently, he became elected as the
first and the fourth president of the Polish Linguistic
Society (19251928 and 19301931). Being at the
peak of his scientific accomplishment Rozwadowski
delivered a lecture before the Societe de Linguistique
de Paris in 1925. He had regarded linguistics as in the
highest degree predestined to formulate generalizations about the mechanisms of human language and
mind; subsequently, he proposed, as the best way to
Rule Borrowing 671
learn more about the relevant features of culture and

the psychical development of man, to undertake the
following three tasks which appeared to him to be
urgent, namely:
1. to describe all linguistic categories and the structures of all languages of the world as constituents
of general grammar;
2. to arrange systematically and notionally all possible lexical entries, derived from all languages of
the world into a general lexicon, with their etymological explanations and semantic values; and
3. to collect products of universal literature, transmitted orally from generation to generation,
as belonging to the literary tradition of mankind
in general.
Dealing with semantics in its relation to grammar,
Rozwadowski argued that semantics is not to be
placed in the same dimension with the disciplines describing other systems of language, such as phonetics,
inflection, derivation, and syntax. The same is not
to be said, as he claimed, with reference to the word
morphology, because every language product, constituting a unity of form and meaning, might be considered either from the point of view of morphology or
from the point of view of semantics.
On the 40th anniversary of his scholarly activity he
was honored by a two-volume Festschrift (Symbolae
grammaticae in honorem Ioannis Rozwadowski.
Krako w, 192728); among its contributors were
M. Rudnicki, J. N. Baudouin de Courtenay,
H. Pedersen, A. Sommerfelt, M. Vasmer, and

J. Vendryes. Important for the history of general linguistics are the following works of Rozwadowski
in particular: Semazjologia, czyli nauka o rozwoju
znaczen wyrazo w. Jej stan obecny, zasady i zadania.
Eos 9, 1903, 17111; Wortbildung und Wortbedeutung, 1904; O zjawiskach i rozwoju jezyka, 1921;
Studia nad nazwami wo d sowianskich, 1948. Most
of his monumental publications have been edited
in the form of Selected writings (Wybor pism, 13,
19501960).
See also: Brugmann, Karl (18491919); Hirt, Hermann
(18651936); Leskien, August (18401916); Pedersen, Holger (18671953); Sievers, Eduard (18501932); Sommerfelt, Alf (18921965); Wundt, Wilhelm (18321920).
Bibliography
Adamska-Saaciak A (1996). Language change in the
works of Kruszewski, Baudouin de Courtenay and
Rozwadowski. Poznan : Motivex.
Man czak W (1960). Zagadnienia ogo lnoje( zykoznawcze
w pracach Jana Rozwadowskiego. In Rozwadowski J M
(ed.) Wybor pism III. Warszawa: Pan stwowe Wydawnictwo Naukowe. 517.
Wasik Z (2001). The development of general linguistics in
(
the history of the language sciences in Poland. In
Koerner E F K & Szwedek A (eds.) Towards a history of
linguistics in Poland. Amsterdam and Philadelphia: John
Benjamins. 351.
Rule Borrowing
S Thomason, University of Michigan,
Ann Arbor, MI, USA
An ongoing debate in historical linguistics revolves

around the general question of borrowability: are
there constraints that bar certain kinds of linguistic
features from being transferred from one language to
another in language contact situations? Historical
linguists are in general agreement that the most common, and most easily identified, interference features
are words. There is also general agreement that new
phonological, morphological, and syntactic rules can
enter a language as a side effect of lexical borrowing.
For instance, the inherited English stress rule was
simple: stress the first syllable in the word. Massive
lexical borrowing from French after the Norman
Conquest of 1066 destroyed this simple rule for the

lexicon as a whole, replacing it, on one account, with
a complex rule that is supposed to account for both
native and borrowed English words: stress the penultimate vowel in the word if the final vowel is lax and
is followed by no more than one consonant, but stress
the last vowel in the word if that vowel is tense and/or
is followed by more than one consonant (Chomsky
and Halle, 1968: 70). But those who believe in
the existence of this rule do not believe that the rule
itself was borrowed from French, because French has
no such stress rule; instead, those who accept(ed)
Chomsky and Halles English stress rule would
argue that it was abstracted by English speakers
from the lexicon as enriched by thousands of French
loanwords.
Similarly, English speakers who pluralize the borrowed noun phrases poet laureate and court martial
Rule Borrowing 671
learn more about the relevant features of culture and

the psychical development of man, to undertake the
following three tasks which appeared to him to be
urgent, namely:
1. to describe all linguistic categories and the structures of all languages of the world as constituents
of general grammar;
2. to arrange systematically and notionally all possible lexical entries, derived from all languages of
the world into a general lexicon, with their etymological explanations and semantic values; and
3. to collect products of universal literature, transmitted orally from generation to generation,
as belonging to the literary tradition of mankind
in general.
Dealing with semantics in its relation to grammar,
Rozwadowski argued that semantics is not to be
placed in the same dimension with the disciplines describing other systems of language, such as phonetics,
inflection, derivation, and syntax. The same is not
to be said, as he claimed, with reference to the word
morphology, because every language product, constituting a unity of form and meaning, might be considered either from the point of view of morphology or
from the point of view of semantics.
On the 40th anniversary of his scholarly activity he
was honored by a two-volume Festschrift (Symbolae
grammaticae in honorem Ioannis Rozwadowski.
Krakow, 192728); among its contributors were
M. Rudnicki, J. N. Baudouin de Courtenay,
H. Pedersen, A. Sommerfelt, M. Vasmer, and

J. Vendryes. Important for the history of general linguistics are the following works of Rozwadowski
in particular: Semazjologia, czyli nauka o rozwoju
znaczen wyrazow. Jej stan obecny, zasady i zadania.
Eos 9, 1903, 17111; Wortbildung und Wortbedeutung, 1904; O zjawiskach i rozwoju jezyka, 1921;
Studia nad nazwami wod sowianskich, 1948. Most
of his monumental publications have been edited
in the form of Selected writings (Wybor pism, 13,
19501960).
See also: Brugmann, Karl (18491919); Hirt, Hermann
(18651936); Leskien, August (18401916); Pedersen, Holger (18671953); Sievers, Eduard (18501932); Sommerfelt, Alf (18921965); Wundt, Wilhelm (18321920).
Bibliography
Adamska-Saaciak A (1996). Language change in the
works of Kruszewski, Baudouin de Courtenay and
Rozwadowski. Poznan: Motivex.
Manczak W (1960). Zagadnienia ogolnoje( zykoznawcze
w pracach Jana Rozwadowskiego. In Rozwadowski J M
(ed.) Wybor pism III. Warszawa: Panstwowe Wydawnictwo Naukowe. 517.
Wasik Z (2001). The development of general linguistics in
(
the history of the language sciences in Poland. In
Koerner E F K & Szwedek A (eds.) Towards a history of
linguistics in Poland. Amsterdam and Philadelphia: John
Benjamins. 351.
Rule Borrowing
S Thomason, University of Michigan,
Ann Arbor, MI, USA
An ongoing debate in historical linguistics revolves

around the general question of borrowability: are
there constraints that bar certain kinds of linguistic
features from being transferred from one language to
another in language contact situations? Historical
linguists are in general agreement that the most common, and most easily identified, interference features
are words. There is also general agreement that new
phonological, morphological, and syntactic rules can
enter a language as a side effect of lexical borrowing.
For instance, the inherited English stress rule was
simple: stress the first syllable in the word. Massive
lexical borrowing from French after the Norman
Conquest of 1066 destroyed this simple rule for the

lexicon as a whole, replacing it, on one account, with
a complex rule that is supposed to account for both
native and borrowed English words: stress the penultimate vowel in the word if the final vowel is lax and
is followed by no more than one consonant, but stress
the last vowel in the word if that vowel is tense and/or
is followed by more than one consonant (Chomsky
and Halle, 1968: 70). But those who believe in
the existence of this rule do not believe that the rule
itself was borrowed from French, because French has
no such stress rule; instead, those who accept(ed)
Chomsky and Halles English stress rule would
argue that it was abstracted by English speakers
from the lexicon as enriched by thousands of French
loanwords.
Similarly, English speakers who pluralize the borrowed noun phrases poet laureate and court martial
672 Rule Borrowing
as poets laureate and courts martial, maintaining the

Noun-Adjective word order of the source language,
are not applying a rule that was borrowed as a rule
from French or Latin; they are merely using a borrowed singular phrase and a plural phrase that was
borrowed separately, even if (for some speakers) such
formations currently have the status of a minor rule.
Occasionally we find analogic extensions that provide evidence for speakers innovative application
of a rule abstracted from borrowed lexical items. So,
for instance, borrowed Latin singular forms such as
alumnus, bacillus, stimulus and the corresponding
borrowed Latin plurals alumni, bacilli, stimuli have
given rise to a minor English rule for plural formation
in foreign-seeming nouns, and this rule has been extended to words which, like octopus, never had such
a plural etymologically. Some English dictionaries
give alternate plural forms octopuses and octopi for
this noun, but not the etymologically expected octopodes. (The word is originally a Greek compound
comprising octo- eight and pous, pod- foot; it does
not contain the Latin noun suffix -us that usually
corresponds to a Latin plural suffix -i.)
For processes like these, there is no controversy.
What is involved is lexical borrowing followed by
the abstraction of a rule from the individually borrowed lexical items, not rule borrowing per se. The
genesis of the rule takes place in English and the
rule is created by English speakers; it does not enter
English as part of the borrowing process. Before
moving on to examples for which a process of rule
borrowing itself might be proposed, we need a definition of rule. For this purpose, the ideal definition
will not be theory-internal, but will instead be maximally inclusive. Trask (1993: 245) offers a definition
that works well: a rule is:
any statement expressing a linguistically significant generalization about the grammatical facts of a particular
language, especially when formulated within the formalism of some particular formal description . . .
This definition covers not only traditional generative

rewrite rules but also lexical generalizations, Optimality Theory constraints, and other generalizations
that typically lack arrows in graphic representations. The topic of this article is thus language
change through the transfer of generalizations from
one language to another.
A slightly less obvious example is found in Asia
Minor Greek, which has borrowed massively from
Turkish in all grammatical subsystems lexicon, phonology, morphology, syntax, and (though the most
comprehensive source of information, Dawkins,
1916, does not provide many examples) lexical semantics. The most-affected Greek dialects have a
vowel-harmony rule that matches the Turkish rule

closely, applying both to Turkish loanwords and to
native Greek vocabulary. This could easily be another
instance of a rule abstracted from borrowed lexicon
and then extended to native words, analogous to (but
more systematic than) the analogically extended
Latin-origin English plurals in -i. But the picture is
complicated by the fact that the vowel harmony rule
is inconsistently applied, even in Turkish loanwords:
the word uva n-us a youth is a Turkish loanword
(though ultimately it comes from Persian), and its
plural in Asia Minor Greek is uva n-iri. This plural
form would violate the Turkish vowel harmony rule,
according to which (among other things) a suffixinitial vowel must agree in the feature [back] with
the last vowel of the stem (Dawkins, 1916: 47, and
see also the discussion in Thomason and Kaufman,
1988: 218). Elsewhere, however, the same dialects
contrast front-vowel and back-vowel plural suffixes
in native vocabulary, e.g., kle fc -iri thieves vs. a rtupuri men, as well as in borrowed Turkish words.
Inconsistencies of this type suggest that bilingual
Greek speakers might possibly have borrowed the
rule itself, independent of the loanwords that would
have provided evidence for it. (The relevant speakers
would presumably have been primarily members of
the Greek speech community, because speakers with
Turkish as their first language were much less likely to
know Greek than vice versa.) Dawkins (1916: 68)
makes an observation that hints at manipulation,
probably unconscious, by bilinguals who regularly
spoke Turkish in addition to Greek:
The fullness with which the vowel harmony is observed
clearly depends on how far the individual speaker is
accustomed to talk Turkish and has the Turkish ear for
these distinctions . . . It [is] more or less prevalent and
thorough in proportion as more or less Turkish is spoken
alongside of the Greek dialect.
A search for cases of rule borrowing turns up all

too many indeterminate examples for which, as with
vowel harmony in Asia Minor Greek, a case could be
made for or against a process of rule borrowing. One
type of indeterminacy is delayed-action syntactic
change that ultimately, but not immediately, results
from the borrowing of a grammatical morpheme.
An especially clear example is the ongoing change
in noun-phrase coordination in Siberian Yupik
(Menovs c ikov, 1969: 128, and see also Thomason
and Kaufman, 1988: 5556). The starting point was
the Yupik suffixal construction, with the comitative
suffix -lju with ([1a] below). Then the conjunction
inkam and was borrowed by Yupik speakers from
the Paleosiberian language Chukchi and used in double-marked coordinate constructions (1b). Finally,
Rule Borrowing 673
the native Yupik suffixes were dropped and NP coordination was expressed solely by the borrowed conjunction (1c) (all three sentences in example 1 mean
on the tundra live birds and beasts).
(1a) NunivaXmi
kijax. taqut
ti$X. iXat-lju
qawaXEt-lju.
birds-COM
on.the.tundra
live
beasts-COM
(1b) NunivaXmi kijax. taqut ti$X. iXat-lju $inkam qawaXEt-lju.
(1c) NunivaXmi kijax. taqut ti$X. iXat
$inkam qawaXEt.
But the historical sequence of events described here

is inferred, not actually attested: Menovs c ikov found
all three constructions in use in Siberian Yupik. The
hypothesized sequence is of a common type, with
double marking here, the comitative suffixes on the
nouns plus the borrowed coordinate conjunction as
a transitional stage between the original construction
and the final construction containing only the
innovative form (regardless of whether the innovation
is expressed by native or borrowed material).
Menovs c ikov does not, unfortunately, say which
speakers used which variants (so that we cannot
know whether, for instance, the oldest speakers
used only constructions like [1a] and the youngest
speakers only constructions like [1c]), but it still
seems reasonable to posit a historical transition
corresponding to (1a)(1c).
The important point in the present context, however, is that the borrowed conjunction arguably
had at first only a minor impact on the languages
syntactic structure, when it was inserted into coordinate NPs as in (1b). The more significant structural
change, abandonment of the comitative suffixes entirely, happened some time after the initial borrowing: that is, even if the historical transition was not
sharply marked by a temporal separation between the
three variants in (1), constructions like (1c) surely
appeared later than constructions like (1b). But in
that case, the end-stage rule NP coordination via
conjunction alone was not itself borrowed from
Chukchi; only the lexical item, with part but not
all of its structural baggage, was borrowed. Even if
Yupik speakers moved to the stage represented by (1c)
as a result of continuing fluent use of Chukchi, we
still cannot prove that the rule itself was borrowed.
Comparable examples are reported from other
contact situations, for instance Pipil (a Uto-Aztecan
language closely related to Nahuatl) in contact with
Spanish in El Salvador, as reported by Harris and
Campbell (1995: 147149, citing Campbell, 1987,
1985). In particular, new relative clause constructions
have been innovated in Pipil that make use of the
relative marker ke, a borrowing from the Spanish
relative marker que. But since the Pipil examples,
like the Siberian Yupik example, begin with lexical
borrowing, they cannot be considered indisputable
instances of rule borrowing.
Another type of indeterminate example is one in

which a rule is lost, not gained, under the influence of
another language that lacks the rule. For instance, just
as the Indo-European language Asia Minor Greek (partially) adopted a vowel harmony rule under the influence of Turkish, so certain dialects of the Turkic
language Uzbek lost their inherited vowel harmony
rule under the influence of Indo-European, specifically
the Iranian language Tadzhik (Menges, 1945; Comrie,
1981: 6566). But although this change obviously
alters the phonological rule system of the relevant
Uzbek dialects, it is not evident that rule deletion is
a plausible candidate for a borrowed rule. Examples
like these, then, will also not provide evidence that rule
borrowing is possible.
Yet another type of contact-induced change that is
problematic for claims of rule borrowing is one in
which, although the receiving language does change
significantly under the influence of the source language, both of the resulting constructions existed in
the receiving language before the source language
influenced it. In such cases, the change is in the frequency and/or the distribution of a pre-existing construction; it is not a case of adopting a completely
new rule. Asia Minor Greek again offers an example.
In the most-affected dialects, the inherited Greek definite article has been almost entirely lost, but it is
retained in the accusative case, the only place in
Turkish grammar in which definiteness is expressed
morphologically (Dawkins, 1916: 46, 87). This dramatic alteration in the morphosyntax of the Greek
article is certainly due to Turkish influence; it cannot
be coincidence that the article has come to be used
always and only where Turkish employs an accusative
case suffix, namely, to mark a definite object. It is also
evident that what has been borrowed into Greek is
the pattern of definiteness marking, because there is
no borrowed Turkish morpheme and Greek retains its
inherited morphosyntactic means of expressing definiteness. Still, from the viewpoint of Greek structure,
the change is in frequency (much reduced) and distribution (greatly restricted) of article usage. Skeptics
who prefer to believe that rule borrowing is impossible will therefore reject examples of this sort on the
grounds that no entirely new category has appeared
in Greek.
A final type of change that is indeterminate from
the perspective of rule borrowing is a process through
which the receiving language does indeed get a new
rule as a result of influence from the source language,
but the new rule is not identical to the sourcelanguage rule. The question here is whether it is reasonable to talk about rule borrowing when (in effect)
only part of a rule has been borrowed. Interference in
Ethiopic Semitic languages from Cushitic languages
674 Rule Borrowing
whose speakers shifted in large numbers to Semitic

provide two relevant instances of morphosyntactic
borrowing. The Semitic languages that were most
heavily influenced by Cushitic have replaced the
inherited Semitic rules for expressing two morphosyntactic categories by Cushitic rules, but only partly.
In both cases, the grammatical morphemes themselves do not change; they were and are native Semitic
affixes. First, the inherited Semitic formation of the
nagative perfect consisted of a verbal prefix; Cushitic,
in sharp contrast, expresses the negative perfect by
means of a verbal suffix. Ethiopic Semitic uses a
prefix-Verb-suffix construction, combining the Semitic prefix with a copy as a suffix. The resulting
construction is neither purely Semitic nor purely
Cushitic. The addition of the suffix to the negative
perfect construction is obviously due to Cushitic
influence (and this analysis is supported by the existence of a wide range of Cushitic features introduced
into Ethiopic Semitic), but the end result is not identical to the equivalent Cushitic construction. Second,
the inherited Semitic formation of the causative also
consists of a verbal prefix. The relevant Cushitic languages form the causative with a doubled verbal suffix. Ethiopic Semitic forms the causative with a
doubled prefix maintaining the placement of the
causative affix but adopting the Cushitic pattern of
doubling the affix.
Examples of this general type, where the end result
of interference (especially in typologically congruent
grammatical subsystems) differs both from the original receiving-language structure and from the
corresponding source-language structure, are easy to
find in phonology. One striking example is the development of a stress pattern, in a dialect of Croatian
spoken near the Hungarian border, that is unique in
all of Serbo-Croatian. According to Pavle Ivic (1964),
the fixed penultimate stress rule of this dialect arose
when Hungarian speakers shifted to the local Croatian dialect and failed to acquire the complex prosodic system, including its free (phonemic) stress,
Instead, they apparently assumed that stress was
fixed, as it is in Hungarian. But, realizing that it was
not fixed on the first syllable, as in the Hungarian
rule, they settled on a kind of average stress placement, on the penult. As in the Ethiopic Semitic example, here too there is no doubt that interference
brought about the change; but the result, though
clearly the adoption of a rule, is not the adoption of
the same rule as in the source language.
All these indeterminate examples provide a rather
clear indication of what sort of example would be a
noncontroversial instance of rule borrowing. The
ideal example would involve no lexical transfer
that is, although there might be lexical transfer from
the same source language to the same receiving language, borrowed morphemes would be unconnected
to the proposed transferred rule. In addition, the
proposed rule borrowing should result in identical
rules in source and receiving language. Fortunately,
the ideal case is not hard to find.
There are two main types of contact-induced structural change that frequently involve little or no lexical
transfer. First, in cases of language shift, the receiving
language is altered as a result of imperfect learning
of its structures by shifting speakers; in such a case,
especially if the shifting group has lower social, economic, and/or political status than the original targetlanguage speech community, lexical transfer may
be minimal. In any case, unless the shifting group
is especially prestigious, phonological and syntactic
interference features will predominate. By contrast,
when imperfect learning is not a factor, i.e., when the
initiators of the changes are fluent bilinguals in the
source and receiving languages, lexical borrowing is
(almost?) always by far the most common type of
interference. (For discussion of this distinction, see
Thomason and Kaufman, 1988; Van Coetsem,
1988; Thomason, 2001: Chap. 4.) Second, in dialect
borrowing, where both lexicon and structure of the
source and receiving languages overlap to a very great
extent, structure is often transferred without morphemes. The same is true of interference between
very closely related languages.
One example in the latter category is an innovative
phonological rule in dialects of certain Kichean languages of the Mayan family that are in intimate contact with Mayan languages of the Mamean branch of
the family. This is a dissimilation rule that palatalizes
velar stops when there is a uvular later in the word
(Campbell, 1998: 74). A morphological example that
also belongs in this category is found in the SerboCroatian dialect of Hvar, as described by Hraste
(1935: 1725). In the 1930s, elderly speakers of the
dialect still used their inherited pattern of syncretism in
the oblique plural noun cases of o-stem nouns, according to which the genitive/locative plural suffix -ih was
opposed to a dative/instrumental plural suffix -ima.
But under the influence of Standard Serbo-Croatian,
younger Hvar speakers had replaced this pattern with
the Standard one, in which the genitive plural suffix is
unique to the genitive and opposed to a single dative/
instrumental/locative plural suffix. But only the distribution of the suffixes changed; the original Hvar
suffixes remained -ih and -ima, in partial contrast
to the Standard Serbo-Croatian genitive plural -a: vs.
-ima. In other words, only the syncretism rule has been
borrowed.
A morphological example resulting from shiftinduced interference led to the emergence of the

Rhetorical Tropes in Political Discourse: Relevant Website

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Rhetorical Tropes in Political Discourse: Relevant Website

Uploaded by

Copyright:

Available Formats

Rhetorical Tropes in Political Discourse 597

Mann W C, Matthiessen C M I M & Thompson S A (1992).

Rhetorical Tropes in Political Discourse

Classical Rhetorical and Actual

rhetorical term encompasses all figures of speech

Rhetorical Tropes in Political Discourse 597

Rhetorical Tropes in Political Discourse

Classical Rhetorical and Actual

rhetorical term encompasses all figures of speech

598 Rhetorical Tropes in Political Discourse

elocution to give glamour to ones speech has widely

Rhetorical Tropes in Political Discourse 599

administration; and the various forms of

Toward a Tropology of the Political

600 Rhetorical Tropes in Political Discourse

Of all tropes, metaphors certainly receive the most

that focus on metaphorical speech in different

Rhetorical Tropes in Political Discourse 601

In addition to personifications, there are many

versus periphery/margin, and of boundary, limit,

602 Rhetorical Tropes in Political Discourse

vehicle/boat/ship: effects of immigration as overcrowded boat; and house/building/door/gateway/

Metonymies (from the Greek: renaming, name

Depending on the relationship between the two

Rhetorical Tropes in Political Discourse 603

the semantic background. Example 9 illustrates how

Synecdoches (from the Greek: understanding one

a queen or a states president; cf. pluralis maestatis

604 Rhetorical Tropes in Political Discourse

will be a representative (active and passive right to

ity in Language; Figures of Speech; Irony; Metonymy;

Bourdieu P (1991). Sozialer Raum und Klassen. Lecon

The entry also lists the idiomatic phrase imperfect

And rhymes that dont involve the last stressed vowel

Even these few familiar examples show, then, that

poetry that a Shakespearean sonnet end with two

In fact, even before getting into any details of such

or may not also create rhymes that are defined in

we come to perceive them are thus fascinating and

The Location of Rhymes

syllable with an unstressed one in what is sometimes

The iambic pentameter of Shakespeares Sonnets

The interest of this for rhyme is that when language

In fact, it seems quite possibly generally true that

In the common case illustrated here, the location

Finally, these possibilities suggest that even

being always an end phenomenon, is not redundant,

Medieval Irish included a variety of stanzaic forms

Some of these internal rhymes, it should be noted,

of Dickinsons poems involve counter-lineation, in

In sum, then, we can say that rhyme is indeed

The Domains of Different Rhyme Types

However, because in English at least, poets mix

Here, too, there are interesting questions about

A second interesting question involves rhymes such

The latter case is, as observed by Kiparsky (1973),

And honored among wagons I was prince

Here, the existence of phonological tiers that

Partial Likenesses in Rhymes

Zwicky describes a form he calls rock rhyme, which

Crucially, Zwicky argues that the rhyme of sleeve

well as in the rhyme traditions of other languages.

of dialect variation over both time and place remains