You are on page 1of 29

http://ccr.sagepub.

com/
Cross-Cultural Research
http://ccr.sagepub.com/content/12/2/117
The online version of this article can be found at:

DOI: 10.1177/106939717701200203
1977 12: 117 Cross-Cultural Research
Ronald P. Rohner
Advantages of the Comparative Method of Anthropology

Published by:
http://www.sagepublications.com
On behalf of:

Society for Cross-Cultural Research


Sponsored by the Human Relations Area Files
can be found at: Cross-Cultural Research Additional services and information for

http://ccr.sagepub.com/cgi/alerts Email Alerts:

http://ccr.sagepub.com/subscriptions Subscriptions:
http://www.sagepub.com/journalsReprints.nav Reprints:

http://www.sagepub.com/journalsPermissions.nav Permissions:

http://ccr.sagepub.com/content/12/2/117.refs.html Citations:

What is This?

- Jan 1, 1977 Version of Record >>


at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
117
Advantages
of the
Comparative
Method of
Anthropology 1
Ronald P. Rohner

Ronald P. Rohner
(Ph.D., Stanford)
is
currently Visiting
Professor of
Anthropology
and Human
Development
and Senior Research
Scientist,
Boys
Town Center for the
Study
of Youth
Development,
at The Catholic
University
of America. He is the author of numerous
journal
articles and
reviews and of several
books, including They
Love
Me, They
Love Me
Not: A Worldwide
Study
of
the
Effects of
Parental
Acceptance
and Re-
jection (HRAF
Press
1975).
Reflecting anthropologys deep idiographic disposition, approximately
80
percent of anthropological
research
falls
on the
idiographic
node
of
the
idiographic-nomothetic
continuum.
Increasing
numbers
of
anthropolo-
gists, however,
are
beginning
to
recognize
the
importance of
cross-cultural,
comparative,
nomothetic research.
Comparativists
use
five principal
meth-
odologies
in their research. These research
designs vary
in their relative
power,
i.e. in their
ability
to eliminate
false hypotheses
about worldwide
relationships.
The
power
as well as the
advantages
and
disadvantages of
each
methodology
are discussed.
[Accepted
for
publication: September 1976.]
In
1896,
Franz Boas wrote what is
perhaps
the most influential
paper
in the
history
of American
anthropology.
His
(1896)
article on &dquo;The Limitations of the
Comparative
Method of
Anthropology&dquo; changed
the course of
ethnological inquiry
from
that time forward. The
paper
struck a fatal blow to the nine-
teenth-century
evolutionists misuse of the
comparative
method,
but it also
prejudiced
most of American
anthropology against
all
forms of
generalizing-bcyond generalizations
based on small-
scale,
regional comparisons.
In
place
of the
comparative
method,
which was associated in the nineteenth
century principally
with
evolutionism,
Boas
proposed
the &dquo;historical
method,&dquo;
and with
this
proposal
he ushered in a
paradigm
of
ethnological
research
that dominated American
anthropology throughout
the next
quarter-century.
As a result of the excesses of the nineteenth
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
118
century
evolutionist
perspective
and,
to a lesser
extent,
of the
twentieth-century
historicalist-diffusionist bent initiated
by
Boas,
comparative
research went into
profound disrepute
within Amer-
ican
anthropology-a
state from which it has not
yet
recovered.
Both Boas and tB1alinowski made
ethnographic
ficldwork the
backbone of
proper anthropological inquiry,
but,
by
its
very
nature,
fieldwork is
particularistic
in its concentration on a
specific problem among
a
specific people,
in a
specific
location,
and in a
specific
time
period. Resulting
from this
emphasis
on
fieldwork,
the
concepts
of &dquo;culture&dquo; and &dquo;cultural relativism&dquo;
assumed the
highest
salience in American
anthropology.
It was
not
long
before the
concept
of cultural relativism was escalated
into &dquo;radical&dquo; cultural
relativism,
a doctrine which states that
every
culture is
different,
therefore
noncomparable.
And it was
not
long
before the allied
concept
of &dquo;functionalism&dquo; was dis-
torted into &dquo;extreme&dquo;
functionalism,
a
dogma asserting
that be-
havior within one
society
cannot be
compared
with behavior in
a second
society.
More
recently,
the
concept
&dquo;emic&dquo; has been
introduced into sociocultural
anthropology.
Some
anthropolo-
gists
have extended this construct into an
&dquo;uncompromising
emicist&dquo;
perspective,
a
position
which states that
only
the con-
ceptual categories recognized by
a native
people
themselves
should be admissible in
anthropological inquiry.
All these &dquo;-isms&dquo; have worked
collectively
to
produce
a
deep
idiographic
bias in American
anthropology (see
Rohner 1975b:
6-17). Idiographers
tend to be
suspicious,
often
antagonistic-
sometimes
rightfully
so-toward the
generalizing
interests of
the few nomotheticists or
generalizers
who have
managed
to
endure
throughout
the last
seventy-five years
of
anthropology.
The
idiographic
bias makes a nomothetic science of man
impos-
sible,
because it denies the
credibility
of
comparisons
and
gen-
eralization across cultural boundaries. The
idiographic
bias not-
withstanding,
however,
a
growing
number of
anthropologists
in
recent
years
have
begun
to
join
nomotheticists in the other be-
havioral sciences in
recognizing
the
importance
of
cross-cultural,
comparative,
and
generalizing
research.
Now,
three-quarters
of
a
century
after Boas landmark
paper
on &dquo;The Limitations of the
Comparative
Method of
Anthropology,&dquo;
it seems
timely
to
present
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
119
an alternative
viewpoint
on the
&dquo;Advantages
of the
Compara-
tive Method of
Anthropology.&dquo;
Any study
that
compares
two or more cultural
systems, parts
of
systems,
or
samples
of individuals within one or more
popula-
tions is a
comparative study.
When this work is done
by
an an-
thropologist
or within the framework of
anthropological inquiry,
however
defined,
it becomes
&dquo;comparative anthropology.&dquo; Using
this
definition,
it is often difficult to
distinguish comparative
anthropology
from
comparative sociology
or from cross-cultural
psychology,
or from the
comparative
interest of
any one
of the
other behavioral sciences. Sometimes the
only
discernible dis-
tinction is that the work is done
by
someone who identifies
himself as an
anthropologist
rather than as a
psychologist,
a
sociologist,
a
political
scientist,
or whatever. This
usage
is neither
whimsical nor
capricious,
but reflects a
growing rapprochement
along
some of the
leading edges
of conventional academic dis-
ciplines.
Research interests in
comparative anthropology range
from
locally descriptive, nongeneralizing (i.e. idiographic)
to world-
wide
generalizing (i.e. nomothetic).
But in all
cases, compara-
tive
anthropology
is characterized
by
the
explicit
and
preferably
systematic study
of similarities and differences
among
two or
more
populations.3
3
Rather than
being mutually
exclusive,
as is often
assumed,
the
idiographic-nomothetic duality actually
forms a continuum
with at least three identifiable nodes:
idiographic, idiographic-
comparative,
and nomothetic. The
concept
&dquo;nomothetic&dquo; is it-
self a matter of
degree,
from
purely
localized
generalizations-
often based on
idiographic comparisons-to
worldwide or
species-
wide
generalizations.
Not all nomothetic
inquiry
is based on
comparison.
Leach
(1961: 1-27)
and L6vi-Strauss
(1963),
for
example,
claim that
general
laws
may
be and should be formu-
lated from a
single
case;
but other
scholars,
especially
those in-
terested in scientific
methodology, argue
that lawfulness is
revealed
only by repetition
or recurrence. The
single
and
per-
haps exceptional
case is fortuitous and not amenable to com-
prehension by
laws. The
truly unique
cannot be treated within
the framework of science. Without
comparison,
one has no
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
120
basis for
distinguishing genuine
causal-functional relations from
phenomena
which are
only coincidentally together ( see
Kbbben
1970:
586).
The kind of
noncomparative,
nomothetic
inquiry
supported by
Levi-Strauss and others is of little concern to
comparative anthropology, except
insofar as conclusions
coming
from this work are
susceptible
to
cross-cultural,
comparativc
verification or falsification.
The
idiographic
node of the
idiographic-nomothetic
continuum
is
particularistic, usually descriptive, noncomparative,
and non-
generalizing.
Each
idiographer
tries to
present
his case in such
a
way
as to
preserve
its
special qualities,
its
particularistic
fea-
tures localized in time and
space.
The unit of
analysis-the
cul-
tural
system,
social
institution,
or the individual-is sometimes
treated as
unique, literally
as a class
by
itself.
Idiographic
re-
search is thus
part
of natural
history,
in that it
attempts
to
establish true statements about
particular events,
processes,
and
situations. As
already noted,
the
idiographic
orientation
repre-
sents a
great part
of the
writing
in
anthropology-probably
the
greatest part-even though
most
anthropologists
do sometimes
talk about
generalizing.4
The
idiographic
orientation
per
se is
of little further interest to
comparativists
or to nomotheticists.
The second
node,
idiographic-comparison, represents
the low-
est level of
inquiry
in
comparative anthropology.
The minimal
intent of
comparison
is to establish likenesses or differences
between or
among
units. A
good
deal of
comparative inquiry
is
also concerned with
generalizing,
but
generalizations
derived
from
idiographic comparisons
hold true
only
for the
specific
units-cultures, institutions,
or individuals-that are
compared.
To what extent one can
generalize beyond
the units
compared
is
entirely problematic.
Strictly speaking,
even
though they
are
nongeneralizing
and
noncomparative (i.e. idiographic), ethnographic
case studies are
by implication
the most
primitive
form of
idiographic compari-
son,
in that
ethnographic description
carries the
implications
of
comparison
between some feature of the &dquo;native&dquo; life studied
and the cultural
system
to which the
ethnographer belongs.
Anthropologists engage
in two classes of
case-study
research.
The
holistic-depictive approach (Bennett
and Thaiss
1967)
is the
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
121
best
known,
at least outside
anthropology.
This is Malinowskian
ethnography,
wherein the fieldworkcr
attempts
to describe in
a holistic manner the culture of &dquo;his&dquo;
people.
The
other,
and
more
common,
form of case
study
is the institutional or
special
topics approach,
wherein the
anthropologist
studies a
specialized
problem among
a
culturally organized population,
for
example
Navaho
kinship, Gurage
nutrition,
Dani
warfare,
or Menomini
acculturation.
Unlike
ethnographic
case
studies,
the controlled
comparison
or concomitant variation
study
is
explicitly comparative.
In these
studies,
the
investigator usually
locates two or more
neighbor-
ing,
but sometimes
distant,
populations
in which one of two
conditions is true:
(1)
relevant variables in the
sample popula-
tions
vary,
but other factors are said to remain
constant;
(2)
relevant variables in the
sample populations
remain
constant,
while all other factors are said to
vary.
Redfields
(1941) study
of four
community types
in Yucatan is a classic
anthropological
illustration of the first
type
of concomitant variation
study.
The
four communities varied in
population
size: one was a
large
city,
one a medium-sized
town,
the third was a
large village,
and the fourth a small
village.
In all other
respects,
these settle-
ments were said to be alike. Redfield then
attempted
to deter-
mine what factors varied
concomitantly
with
population
size.
The work of
Murphy
and Steward
(1956)
illustrates the sec-
ond
type
of concomitant variation
study,
where relevant vari-
ables remain constant while others
vary. Murphy
and Steward
did a controlled
comparison
of
parallel processes
of accultura-
tion
among
the Mundurucu of Brazil and the
Montagnais
in
subarctic North America. The two societies were alike in that
they represented
the same level of sociocultural
integration-
i.e. both were
politically
autonomous local units.
Moreover,
both were
exposed
to similar acculturative
pressures, namely
to outside commercial influences. In other
respects,
the two so-
cieties were
quite
different. The
Mundurucu,
for
example,
were
tropical
forest hunters and
horticulturists,
living
in
semiperma-
nent
villages;
whereas the
Montagnais
were
subarctic,
nomadic
hunters of
large migratory game. Murphy
and Steward com-
pared
these two
societies,
showing
that even
though they
were
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
122
different in most
respects-except
for the critical condition of
being
at the same level of sociocultural
integration-they
re-
sponded
in similar
ways
to similar acculturativc
pressures.
Regional
trait distribution
studies,
including
the
idiographic
statistical method
pioneered by
Boas
(1894, 1895)
in his work
on Northwest Coast folk
tales,
form a second class of true idio-
graphic-comparative inquiry.
These studies involve
collecting
data on a list of traits from all the tribes of a
region,
or even
within a
continent,
and then
constructing
an intercorrelation
matrix,
in which each trait is correlated with all other traits.
From a cluster
analysis
or factor
analysis
of this
matrix,
it is
possible
to discern
geographic regions
where
groupings
of traits
-i.e. trait
complexes―cluster.
In the
history
of
anthropology,
one of the
important goals
of this method has been to
help
reconstruct,
at least
inferentially,
the culture
history
of a
specific
region
or of certain traits or tribes. Driver and
Masseys (1957)
&dquo;Comparative
Studies of North American Indians&dquo; is one of the
most extensive
applications
of this method.--
Nomothetic
comparison
is the third node on the
idiographic-
nomothetic continuum. As
already
indicated,
the extensiveness
of
comparison
and of
generalization
is variable within the nomo-
thetic-comparison range
of the continuum.
Nomothetic-compara-
tive
inquiry
sometimes
yields purely
localized
generalizations
that are
only indistinctly
different from the conclusions of the
basically nongeneralizing, idiographic-comparative
node. At the
other
extreme,
some
nomothetic-comparative
research strives to
establish
lawlike, worldwide,
universalist
generalizations,
or
principles
of human behavior.
The
category
of
nomothetic-comparative inquiry
contains two
distinguishable,
but sometimes
related,
classes of research. The
first,
hologeistic
research,
or cross-cultural
survey research,
as
it is often
called,
is oriented toward worldwide correlational
tests of
hypotheses
or theories. The
hologeistic
method
employs
a worldwide
sample
of societies from the
ethnographic
record
of
anthropology
to
statistically
test
hypotheses
about the trans-
cultural
relationship
between two or more
psychological,
social,
or cultural variables. Naroll
(1973: 354-55)
identifies five sub-
types
of
hologeistic
studies: a lwlocllltllral
study
is a
&dquo;hologeistic
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
123
study
which draws
upon all,
or
substantially
all,
human cul-
tures known in the last four
centuries&dquo;;
a llOlonational
study
is
a
&dquo;hologeistic study
which draws
exclusively
on modern
nations,&dquo;
a liolohistorical
study
is a
&dquo;hologeistic study
which draws ex-
clusively
on
preindustrialized,
traditional,
record-keeping
so-
cieties,&dquo;
a
holoarcheological study
is a
&dquo;hologeistic study
which
draws on the
archeological
remains of
nonrecord-keeping
so-
cieties,&dquo;
and a holoethnohistorical
study
is a
&dquo;hologeistic study
which draws on the records made
by
outside observers of the
history
of a
nonrecord-keeping society.&dquo;6
More than 425 holo-
cultural studies and
nearly
600 holonational studies have been
completed
so far
(see
Naroll,
OLeary,
and
Sigelman 1974),
but
very
few holohistorical studies arc in
print.
As of
July
1974,
only
one
holoracheological study
had been
done,
and
only
one
holoethnohistorical
study
had been
reported (Naroll, OLeary,
and
Sigelman
1974:
3).
Holocultural and holonational methods
have been used to
investigate
a wide
range
of
issues,
including
kinship
and
marriage,
cultural
evolution,
parent-child
relations,
personality development,
deviant
behavior,
economic
activities,
political
structure,
and
religion.
The universalist
approach
is the second class of research
within the
category
of
nomothetic-comparative inquiry.
The
universalist
approach
is a multimethod and
usually
interdisci-
plinary
research
strategy
that has as its
objective
the establish-
ment of
scientifically
derived
generalizations
or
principles
about
human
behavior-generalizations
that are
specieswide
in their
applicability (Rohner 1975b). (The
central characteristics of
the universalist
approach
are detailed
later.)
Universalist
prin-
ciples
are of two basic
types: universally
valid,
descriptive
state-
ments and
universal,
causal-functional
relationships.
These two
types
are described in
greater
detail below. A
good
deal of
my
own research
( Rohner
1970, 1975a, 1975b, n.d., 1976; Rohner,
DeWalt,
and Ness
1973;
Rohner and Katz
1970)
on the world-
wide antecedents and
consequences
of
parental acceptance
and
rejection
falls within the
category
of universalist research.
So,
too,
does the work of some other
anthropologists,
such as
Berlin,
who is concerned with universal semantic
processes
in
color lexicons and in
cthnobiological
lexicons
(Berlin
and
Kay
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
124
1969).
Researchers outside
anthropology
have also
pursued
uni-
versalist research. Some
principal
concerns here include the de-
velopment
and structure of human
cognitive processes (see,
for
example, Piaget
1970;
Dasen
1972),
the evolution of hominid
social bonds
(see,
for
example, Bowlby
1969;
Freeman
1966;
Freedman
1968;
Hamburg 1968),
the
purported
instinctual basis
of human
aggression (see,
for
example,
Lorenz
1966; Ardrey
1966;
Morris
1967;
Freud
1930),
the evolution of sex differences
in behavior
(see,
for
example, Goldberg
1973;
Freedman
1968;
Rohner
1976;
Tiger 1969),
the evolution of a universal culturc
pattern (see,
for
example,
Sebeok
1968;
Keesing
1973;
Fox
1970),
and a
universalist-evolutionary
orientation to
linguistics (see,
for
example, Chomsky
1965;
Greenberg
1975;
Jakobson
1968;
Lenneberg
1964;
Osgood
et al.
1975).
All of these researchers
have in common an interest in the evolution of human
social,
emotional,
or
cognitive
behavior
( as distinguished
from human
morphology).
Not all universalist
research, however,
is construed
from an
evolutionary, biological,
or
ethnological point
of
view,
although
it is
probably
true that at some
point
in all such re-
search the facts of
biology, genetics,
or evolution must be
recognized.
Worldwide,
generalizing
research in the
nomothetic-compara-
tive mode
represents
but a small
percentage
of the work done
in sociocultural
anthropology today.
In
fact,
Table 1
clearly
exposes anthropologys idiographic
bias.
Eighty percent
of an-
thropological inquiry
in
1973, 1974,
and
1975,
as measured in
three
leading anthropology journals, employs
the
idiographic
case-study approach-mainly
of the institutional or
special topics
variety. Comparative
research,
as defined
by
the
idiographic-
comparative
and the
nomothetic-comparative modes, represents
only
about 20
percent
of the work done
currently
in
general
anthropology.
True
nomothetic-comparative inquiry
is
only
about
one-fourth of that.
Table 1 does not mirror
accurately
the absolute number of
hologeistic
studies
appearing regularly
in behavioral science
periodicals,
because the
majority
of these studies are
published
in
journals
other than the American
Anthropologist,
the Amer-
ican
Ethnologist,
and
Ethnology,
from which the data in Table
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
125
Table 1. Classes
of
Research in Sociocultural
Anthropology,
1973
Through
1975
Data in Table 1 are based on 207 articles
published
in the America
Anthropologist,
American
Etl11101ogist,
and
Ethnology
from 1973
through
1975.
Evelyn
C. Rohner coded 100 of the articles in the American Anthro-
I)ologi.st
and in
Ethnology published during
1973 and
1974;
Patricia A.
Jehle
coded 72 articles in the American
Etlitiologist,
from its
inception
in
1974
through 1975,
and 35 articles in the American
Anthropologist
and
in
Ethnology published
in 1975.
Only
data-oriented
papers
derivative in
some
way
from fieldwork were tabulated. Theoretical articles were
omitted,
as were
papers
outside sociocultural
anthropology.
1 were obtained.
Nonetheless,
the table does reflect the fact
that the
proportion
of
hologeistic
and universalist research to
the conventional and dominant research
paradigm
of
general
sociocultural
anthropology
is still
definitely
in a
minority position.
Data collected
by
Lewis
(1966)
in 1956 from an examination
of 220
journal articles, dissertations,
and books show
that,
in
some
respects,
the
anthropological zeitgeist
has
changed
little
in the
past twenty years.
For
example
Table 1 shows that al-
most 95
percent
of
anthropological
research
today
is either
idiographic
or
idiographic-comparative,
but not nomothetic-com-
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
126
parative.
Lewis found that about 75
percent
of all
comparative,
studies from 1950
through
1954 were conducted within a
single
continent, nation,
culture
area,
or local cultural
group,
thus
showing
a historical
continuity extending
to the
present
in an-
thropologys
s
preferencc
for small-scale
comparison (i.e.
idio-
graphic-comparison )
rather than
large-scale comparisons (i.e.
nomothetic
comparison) anthropologists
make
any sys-
tematic
comparisons
at all. Lewis data also show that about 13
percent
of the research done in
comparative anthropology
was
essentially hologcistic,
or what he called &dquo;random or
global
comparisons&dquo;
based on
library
research
( Lewis
1966:
61).
The
intervening twenty years
have shown a
significant change
in the
incidence of
hologeistic
research. Now 26
percent
of the
pub-
lished work in
comparatice anthropology
is
hologeistic
as mea-
sured in Table
1,
but the table underestimates
dramatically
the
true contribution of the
hologeistic
method to
comparative
anthropology.
I now want to rotate
by
45 the axis of
my argument
and
look at the
methodologies employed by anthropologists doing
comparative
research,
a
topic
that cross-cuts both modes of
comparative
research,
i.e.
idiographic-comparison
and nomo-
thetic-comparison.
Before
continuing,
however,
I need to
amplify
the
concept
of
methodology
as I use it here.
Mcthodology
refers to dis-
tinguishable classes, traditions,
or
paradigms
of
research,
each
with its own natural
history, employing
a
specific logic
and
basic
assumptions
or
epistomology,
and each
comprising
one
or more discrete methods
(i.e.
research
procedures),
such as
questionnaires,
field
schedules,
interview
schedules,
or behavior
observations.
Every methodology
has certain
advantages
and
disadvantages, strengths
and
weaknesses,
as well as the
poten-
tial for certain kinds of bias. Each
yields
certain kinds of in-
formation but does not
yield
other kinds.
Just
as the kind of
fish one catches
depends
on the net that is
used,
and the net
that
is
employed
is
designed
to catch the fish one
values,
so
the kind of data one collects
depends
on his
methodology,
and
the
methodology
is selected to
gather
the kind of data one
values.
Experimental design, sample survey research,
the holo-
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
127
geistic
method,
and
participant
observation field research are
all illustrations of different
methodologies.
One of the salient characteristics of
every methodology
is its
relative
power.
The
power
of a
methodology
in
comparative
research is defined as its
ability
to eliminate false
hypotheses
about human behavior in trorldtcide
perspective.
The
greater
the
capability
of a
methodology
to eliminate a false worldwide
hypothesis,
the
greater
is its
power.
The
concept power
as it
is
applied
to
comparative
research
methodology
is also familiar
in statistical
inference,
in a somewhat more delimited sense.
There,
power
refers to
Type
II or Beta
error,
i.e. the risk of
failing
to
reject
an alternative
hypothesis (vs.
the null
hypothe-
sis)
when the
hypothesis
is
actually
false. The
power
of a
methodology
is variable. Some
methodologies
have low
power,
and others have
increasingly higher
levels of
power.
The re-
mainder of this
paper
reviews a rank
ordering
of the
power
of
the
principal methodologies
or
approaches
used in
comparative
anthropology.
Table 2 sets out
graphically
the
power-ranking
of the
major approaches
in
comparative inquiry
in sociocultural
anthropology.
As indicated
earlier,
the case
study approach
is the most
primitive methodology
used in
comparative anthropology.7
Since
the case
study approach
is
essentially noncomparative-or
com-
parative only by implication-one
can seldom tell how
repre-
sentative of some
larger population
an individual case
really
is.
Therefore,
within the context of a
probability
model,
the case
study
method has
low-virtually zero-power.8
Since the
prob-
ability
model assumes as a matter of course that
exceptions
will
occur,
a case that fails to
support
an
hypothesis
cannot do
more than make one wonder about the
veracity
of the
hypothe-
Table 2. Poicer
Ranking of Approaches
to
Comparative Anthropology
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
128
sis,
or about the
representativeness
of the case. On the other
hand,
within the context of a mechanical
model,
where a
single
exception
to a
theory
or
hypothesis
is sufficient to discredit that
theory,
the
casc-study
method has
great power,
in that a
single
deviant case falsifies a
theory.
Few mechanical models have
been
developed recently
in
comparative anthropology,
but
many
were advanced in the
history
of behavioral science. G. S. Hall
(1904),
for
example, postulated
the worldwide invariance of
adolescent
stress,
and Freud
(1950) postulated
the invariance
of the
Oedipus complex.
In both
instances,
critical
ethnographic
cases were marshaled
by
other scholars to discredit the idea of
specieswide
invariance
(see,
for
example,
Mead 1928 and Mal-
inowski
1927,
respectively ) .
Since mechanical models are rare in
comparative anthropology,
the remainder of this
paper
concen-
trates on
methodologies
for
idiographic-comparisons
and nomo-
thetic-comparisons
within the context of a
probability
model.
The controlled
comparison,
or concomitant variation
study,
and the
regional
trait-distribution
study
both fall within the
idiographic-comparison
sector of the
idiographic-nomothetic
con-
tinuum.
Accordingly, they
both have low
power.
In
addition,
some concomitant variation studies and some
regional
trait-
distribution studies-for
example
the
regional study
interested
in
establishing
taxonomies
(e.g.
Driver
1961;
see Naroll 1968:
242 )-are basically nongeneralizing
but nonetheless
comparative.
This form of
comparative study
has
very
low
power,
even
though
it
may
be useful for
discovering
new variations in be-
havior and it
may
also
help
to delineate
working hypotheses
about the interaction between variables. Other concomitant vari-
ation and
regional
trait-distribution studies are at least min-
imally generalizing
in intent. To this extent
they
share charac-
teristics with the
nomothetic-comparison node,
and
accordingly
have somewhat
higher
but still
seriously
restricted
power. They
have an
advantage
over the
idiographic methodologies
described
up
to
now,
in
being
able to establish correlations within or
among
the various units
compared,
but
rarely
can
generaliza-
tions be extended
confidently beyond
these bounds.
Concomitant variation studies are
frequently
based on
paired
comparisons,
but a
comparison
of two natural cases cannot be
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
129
used for
hypothesis testing.
In
fact,
such a
comparison
is un-
interpretable,
even
though
it
may
be in other
respects
a fruitful
source of
insight (cf. Campbell 1961).
The concomitant varia-
tion
approach
also has other weaknesses
contributing
to its
low
power.
One of the most serious of these is the
inability
of
the
methodology
to
distinguish idiographic
from worldwide no-
mothetic correlations. The
question
here is &dquo;whether the cor-
relations observed
might
not reflect artifacts of culture-trait
borrowing
from a common source rather than fundamental ten-
dencies in the nature of
society
or culture&dquo;
(Naroll
1968:
241).
Regional
trait-distribution studies are
usually
based on more
systematic
and formal data collection
procedures
and are to
this extent more sound
methodologically
than concomitant varia-
tion studies. But trait-distribution
studies, too,
have the
problem
of
distinguishing regionally-specific
correlations from worldwide
causal-functional
relationships.
Thus,
the
regional
trait-distribu-
tion
methodology
is unable to
falsify unambiguously
a world-
wide
hypothesis.
In other
respects,
it shares much in common
with the
hologeistic
method.
With the
hologeistic
method,
comparative anthropology jumps
to the
nomothetic-comparative
sector of the
continuum,
and
from low
power
to
high power.
In
fact,
the
hologeistic
method
is the most
powerful, single methodology yet developed
for
worldwide
hypothesis-testing
research. The method is also the
most
complex single methodology
that is used in worldwide
comparative
research? As indicated
earlier,
the
hologeistic
meth-
od draws on a worldwide
sample
of societies for the
purpose
of
statistically testing hypotheses
about the transcultural rela-
tionship
between two or more variables. Insufficient attention
to five critical
problems
often mars the results of
hologeistic
research.
Among
these issues are: the
problem
of
sample
selec-
tion
(including questions
about the
sampling
universe,
sampling
units,
sampling procedures,
and tests for coder
bias);
definition
and
operationalization
of
variables;
tests for
ethnographer
bias
or data
quality
control;
and
hypothesis
formulation and
testing
(including
statistical
procedures
and other matters of research
design
related to data
analysis-e.g.
the
&dquo;mudsticking&dquo; problem,
regional testing
of
hypotheses,
and
causality
of
correlations).
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
130
In the remainder of this discussion of the
hologeistic
method,
I will concentrate on the holocultural
method,
because this is
the most common
type
of
hologeistic study published.
Eth-
nographies
on which the holocultural method is based describe
the
typical
behavior of a
population
of
people.
The method
thus measures
regularities
in standardized or
customary
behavior
within total communities the world over. For this
reason,
the
holocultural method is
outstanding
for
distinguishing culturally
conditioned from universal causal-functional
relationships.
To
this
extent,
the holocultural method has
high power,
but even
when used
properly-which happens infrequently-it
is not the
most
powerful methodology
for
worldwide, comparative
research.
One of the reasons for this
qualification
is that it
has,
like
any
single methodology,
certain limitations for
worldwide,
compara-
tive,
and
generalizing
research. Its
great strength
for
measuring
the
regularities
in behavior within total communities is at the
same time a
limitation,
in that the
methodology
can
give
no
information about intracultural
variability
in behavior of indi-
viduals.
The universalist
approach
is at the extreme nomothetic end
of the
idiographic-nomothetic
continuum,
and it is also the
most
powerful methodology
used in
comparative anthropology-
or,
for that
matter,
in
any
of the other
comparative
behavioral
sciences. Because it is a multimethod research
strategy,
it is not
marred
by
the weaknesses of a
single methodology.
In
fact,
it
controls for the weaknesses of each
component methodology
through
a
process
called the
triangulation
or
convergence
of
methodologies,
which will be described later.
By
its basic
design,
the universalist
approach attempts
to establish
scientifically
de-
rived
generalizations
about human
behavior,
i.e.
principles
of
behavior that are
specieswide
in their
applicability.
Generali-
zations
coming
from the universalist
methodology may
take the
form either of universal causal-functional
relationships
or of
universally
valid
descriptive
statements. The latter form of
gen-
eralization is not concerned
simply
with the
theoretically
bland
cultural universals that
anthropologists
describe in
every
intro-
ductory
textbook-for
example
&dquo;all societies have a form of sub-
sistence
economy,&dquo;
or the
&dquo;family
in one form or another is
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
131
universal,&dquo;
or an &dquo;incest taboo is found
everywhere,&dquo;
or &dquo;all
people
have some form of
religious
belief.&dquo;
Rather,
the univer-
salist
approach
is more interested in
comprehending
the mech-
anisms or
processes creating specific universals-e.g. Why
is an
incest taboo universal? Within this level of
generalization,
the
universalist
approach
asks not
simply
&dquo;~Vhat are the
major
descriptive
facts of human life?&dquo; but
&dquo;Why
are these
descriptive
facts true?&dquo;
Universal causal-functional
relationships,
the second level of
generalization,
deal with a somewhat different issue. Here the
concern is with worldwide
relationships among phenomena.
These
relationships may
be causal in the sense that one variable
produces
a commensurate
change
in a second
variable,
or
they
may
be functional in the sense that a
change
in one variable is
associated-but
perhaps only indirectly-with
a commensurate
change
in a second variable. In all forms of the universalist
ap-
proach,
however,
scholars want to be sure that the discovered
relationships
are valid for all mankind-either within all human
populations
or
among
all
populations
where certain
limiting
conditions exist. In other
words,
we must be assured that these
relationships
are not
simply
an artifact or
special
circumstance
of the
unique setting (e.g. specific
culture,
gene pool,
local
region)
in which the
relationship
was discovered.
The universalist
approach comprises
two
complementary
facets
-a
conceptual
facet and a
methodological
facet.
Conceptually,
and at the
highest
level of
abstraction,
the universalist
approach
asks the
philosophically
based
question:
&dquo;What does it mean
to be a human
being?&dquo;
The universalist
approach
asks about
the nature of human nature
or,
more
specifically,
about re-
searchable
(i.e. operationalizable)
features of &dquo;human nature.&dquo;
From this
point
of
view, then,
it should be clear that the uni-
versalist
approach
is not interested
simply
in the behavior of
White
Canadians,
Black
Americans,
Kwakiutl
Indians,
or Turkish
peasants-or
even about a
comparison
between
any
two or more
of these
groupings-but
rather in mankind as a whole. In order
to do universalist
research,
scholars must make basic
assump-
tions about the nature of man.
Principally they
must make a
supposition comparable
to the one
anthropologists
call the
&dquo;psy-
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
132
chic
unity
of man.&dquo;
They
must assume that all normal
(i.e.
non-
pathological)
humans are
subject
to the same
developmental
tendencies
and,
additionally,
that at birth all normal humans
share the same
general capacities
for
thought, feeling,
and action.
This
emphasis
on universals does not mean that the uni-
versalist
approach disregards significant cognitive,
emotional,
or other behavioral differences
among
human
populations.
In-
deed,
an interest in human
variability
is the
complement
of
the search for universals. For
example
the
thought processes
of
young
children
may prove
to be
qualitatively
different in
some
respects
from the
cognitive processes
of
adults,
but the
cognitive functioning
of children at a certain
age may
turn
out to be more or less invariant
throughout
our
species.
If this
is
true,
these
developmental
likenesses and differences would
be
especially interesting
to the universalist
approach.
Unlike the
idiographic
or
particularistic
orientation,
which often focuses
on
superficial
local differences
(e.g.
local &dquo;cultural&dquo;
differences),
the universalist interest in
variability
relates to
significant
differ-
ences in the behavior
repertoires
of
subpopulations
within our
total
species.
Methodologically,
the universalist
approach
does not dictate
any given procedure
or class of research
techniques.
Indeed,
quite
different methods are
appropriate
for
asking
different
kinds of
questions.
The universalist
approach
does
presume,
however,
the
presence
of at least minimum standards of scien-
tific
inquiry.
The
generalizations
that are to be elevated to the
level of universalist
principles
must be able to withstand sci-
entific
scrutiny: they
must be
supported by empirical
evidence
that was collected in an
objective
and
impersonal manner,
and
the
proccdures
used must be
open
to
public
review and
replica-
tion,
thereby assuring
that the
purported principles
are
capable
of verification or falsification
by independent investigators.
In
addition,
investigators
must
employ only
variables and
proce-
dures that are
transculturally equivalent,
thus
providing
assur-
ance that the variables can be measured
directly by
the same
procedures
in different
societies,
and that the measures can be
compared (see
Sears
1961;
Brislin et al. 1973:
13-14, 24-29).
These minimal
requirements
are needed in order to
distinguish
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
133
appealing,
but nonetheless
speculative, specieswide generaliza-
tions from
scientifically-derived generalizations.
An additional
point regarding
the
methodological
and
concep-
tual mix
comprising
the universalist
approach
is the need for
comparative
research. Insofar as behavioral scientists are inter-
ested in
establishing
valid,
specieswide generalizations
about
man,
then their research
design
must consider relevant varia-
tions found
throughout
the
specics-or perhaps simply
find out
whether
any
variation exists with
respect
to the behavior in
question.
In either
case,
the
investigator
will be led to some-
thing
like a
comparative,
worldwide
sampling design
in his re-
search. The rationale here is that if behavioral scientists want
to be confident about the
specieswide generalizability
of their
principles, they
are
obligated
to show that the
principles
can
indeed be
generalized beyond
the
population
or
cognate popu-
lations from which
they
were
originally
derived or discovered.
This
point
has been
emphatically repeated
on numerous occa-
sions both in
anthropology
and
elsewhere,
especially
in cross-
cultural
psychology (see,
for
example,
Brislin et al. 1973:
143-44;
Dawson 1971:
291;
DeVos and
Hippler
1969:
324;
Jahoda
1970:
58-59 ) .
The
past
decade has witnessed a
dramatically mounting
in-
terest in worldwide or
specieswide generalizations. Many
of
these
generalizations
are
speculative,
but others
approach
the
standards of scientific
credibility
and are thus consistent with
at least this
requirement
of the universalist
approach.
Wide-
spread
interest in
scientifically
derived
generalizations
about
human
behavior,
especially
in universal causal-functional rela-
tionships,
has been stimulated in
comparative
behavioral sci-
ence more
by
the
hologeistic
method than
by any
other
single
influence. In
fact,
by
its
very
nature,
as I have
already pointed
out,
the
logic
of the
hologcistic
method is oriented toward uni-
versalist
principles.
Some
hologeistic
studies more
closely ap-
proximate
the
requirements
of the universalist
approach
than
any
other line of
inquiry developed
so
far,
but these studies
are not sufficient
by
themselves to establish
unequivocably
the
specieswide principles
of human behavior for which we are
searching.
I turn now to the illustration of a multimethod re-
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
134
search
strategy
that is sufficient for
establishing
these univer-
salist
principles.
As I
pointed
out
earlier,
successful universalist efforts must
not
only
be
generalizing
in intent but must also conform to the
minimum standards of
science,
deal with
transculturally equiv-
alent variables and
procedures,
and at some
point
test their
principal hypotheses
in an
adequate pancultural sample.
These
are but minimum standards for the universalist
approach.
In
addition,
as
already
noted,
researchers should
recognize
that
every &dquo;methodology&dquo;
and
every specific
research
procedure (or,
more
simply, &dquo;method&dquo;)
has certain
strengths
and
weaknesses;
each will
give
certain kinds of information and not
others,
and
each has the
potential
built into it for
systematic
error or bias.
This
being
true,
it is
possible
for research results to reflect this
method bias rather than to be a true measure of the behavior
we wish to know about.
Any
such results are
disastrous,
no less
for the universalist
approach
than for
any
othcr kind of research.
To avoid the
possibility
of
interpreting
&dquo;method bias&dquo; as
being
a true measure of the behavior under
question,
serious univer-
salist researchers
triangulate ( see Figure 1, below )
their results
whenever
possible by employing
a multimethod research strat-
egy. They employ
two or three
independent methodologies
in
order to determine the extent to which the same conclusions
emerge
when
multiple
and
independent
measurement
processes
are used-none of which shares the same weaknesses or
poten-
tial for bias. Our confidence is increased insofar as we
get
con-
verging
results from the use of different
tests,
especially
when
these tests
are
performed
within the framework of
separate
methodologies.
All these
methodological
and
conceptual guidelines
are incor-
porated
into the research
design directing my
own research on
the worldwide causes and
consequences
of
parental acceptance
and
rejection.
In this
work,
we ask the
question,
&dquo;Do human
beings
the world over
respond
the same
way
to
parental accep-
tance
(or rejection), regardless
of cultural
context,
physical
type,
environmental
conditions,
or other
limiting
conditions?&dquo;
We also
ask,
&dquo;Do
parents everywhere
who live under
specified
conditions tend to interact with their children in the same
way
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
135
with
respect
to
warmth, hostility,
and
neglect?&dquo;
These are uni-
versalist or
specieswide questions-questions
that,
when
appro-
priately
answered,
may
be elevated to the level of universal
&dquo;principles&dquo;
of human behavior.
In order to deal with an issue as
complex
as this
one,
we
employ
a research
strategy
that
incorporates
three distinct &dquo;tra-
ditions&dquo; of research or
methodological components (diagramed
in
Figure 1), namely,
the cross-cultural
survey component,
the
psychological
research
component,
and the intracultural com-
munity study component. Figure
1 shows that the
methodologies
used in this research
produce overlapping
results. It is in the
hatched
area,
where all three
methodologies converge,
that
the
results have
successfully
survived the
onslaught
of the multi-
method research
strategy,
and it is in this area that the uni-
versalist or
specieswide principles
are to be found.
The most
highly developed component
of this work is the
one
using
the cross-cultural
survey (i.e. holocultural)
method,
wherein we
employ
a worldwide
sample
of 101 communities
(i.e. &dquo;societies&dquo;) representing
a stratified
sample
of the worlds
known and
adequately
described cultural
systems.
The
objec-
tive here is to test
statistically hypotheses
about the worldwide
relationship
between
parental acceptance-rejection
on the one
Figure
1.
Triangulatiol
of
M etlwdologies
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
136
hand and
specific personality
characteristics of children and
adults,
expressive
features of
society,
and societal maintenance
systems
on the other. The
goal
of this
component,
as for the
project
as a
whole,
is to
develop
a
theory
which will allow us
to
reliably predict
the worldwide
consequences
of
parental
ac-
ceptance-rejection
for
personality development
and
functioning,
as well as for selected
expressive
features of
society,
and also
to
predict specific
environmental and
maintenance-system
con-
ditions under which
parents everywhere accept
or
reject
their
children.
The second
component,
the
psychological
research
component,
involves research in North America on various
aspects
of the
rejection-acceptance problem, including
child abuse. This com-
ponent
also
incorporates cross-species
research on
aspects
of
developmental psychobiology, ethological studies,
and other dis-
ciplines relating
to the matenial behavior of mammals
(see,
for
example, Denenberg
1969: DeVore
1963;
Harlow
1971),
and it includes an intensive and
systematic survey
of a vast ex-
perimental
and observational literature in the
psychological
sci-
ences on the causes and
consequences
of
acceptance-rejection.
Psychological
research
complements
cross-cultural
survey
re-
search,
in that the former deals with interindividual
variability,
whereas holocultural research deals with intercultural
variability.
It is often
possible
to
experimentally manipulate
and control
psychological
variables within North America in
ways
that can-
not be done in cross-cultural
research;
the holocultural method
is nonetheless
indispensable, because,
among
other
things,
it
lets
investigators
measure the extent to which
psychological
re-
search done in North America can be
generalized
to the entire
world.
Psychological
research within North America cannot
by
itself
distinguish culturally dependent
from
specieswide
devel-
opmental
tendencies. In
fact,
the vast
portion
of all
psychologi-
cal research is in one sense
culture-bound,
in that it has been
done within the United States. More
particularly,
it has been
conducted on a minuscule and
unrepresentative
sample
of the
American
population-i.e.
on
college students,
especially
intro-
ductory psychology
students
(Carlson
1971:
204).
As a
result,
many psychologists
and
anthropologists
have raised serious
ques-
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
137
tions about the
pancultural generalizability
of a
significant por-
tion of
psychological
research.
Intracultural
community
studies,
or case
studies,
form the
third
component
of this research. The
community study ap-
proach
involves
long-term anthropological
and
psychological
field
investigations
of communities within the United
States,
but
especially
in culture areas outside North America. The
gen-
eral cultural
setting
as well as
parent-child relationships
are
studied within each
community,
and
personality
assessments are
made of the
sample
children and their
parents.10 Community
studies
provide
invaluable information about the influence of
natural
(e.g.
cultural, climatic,
etc.) settings
on behavior and
personality development.
The
community study (i.e.
case
study)
method also
provides
the
opportunity
to
vary systematically
cul-
tural and other social and environmental
conditions,
while si-
multaneously allowing
for the measurement of individual vari-
ability
as well as behavioral uniformities within each
community.
This
component,
like the
psychological
research
component,
employs
an
interdisciplinary
and multimethod research strat-
egy.ll
But unlike the other two
methodologies,
the
community
study component
concentrates on
tcithil1-community
consistencies
and
variability
in behavior. In
every community
some
parents
are warmer or more
accepting
than
others,
even
though
the
general
cultural nomi
may
tend toward
rejection.
Similar varia-
tions in other relevant forms of behavior are found within all
communities as well. The universalist
theory
in our research
postulates
that
rejection (or acceptance) by
itself is
sufficient,
but not
necessary,
to
produce
certain
specified consequences,
regardless
of the &dquo;culture&dquo; of a
people,
their
physical type,
or
any
other conditions that
might
alter their behavior.
Thus,
to
take an extreme
example,
the
theory predicts
that an
accepted
child within
any given family
will
develop
certain character-
istics that are more like the characteristics of
accepted
children
elsewhere-perhaps
under
widely
different cultural and other
conditions-than the characteristics of his own
rejected sibling.
The
community study component
and
aspects
of the
psycho-
logical
research
component provide
the
opportunity
to
study
in
situ
many
of the
relationships
found in cross-cultural
surveys.
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
138
This
technique
of
subsystem rcplication (i.e. testing
within a
single
cultural
system
the results reached in holocultural re-
search )
contributes to the assessment of the
validity
and
general-
izability
of results
coming
from the other two
methodologies.
The
universalist, multimethod,
research
strategy
described
here is
likely
to involve
investigators
in
interdisciplinary
research
-as it does in
my
work-because,
it
seems,
most of the behav-
ioral sciences have
only
one or a
very
few
methodologies (i.e.
general
classes or
paradigms
of
research)
that are
regarded
as
being appropriate
for
answering
the
questions customarily
asked
within each
discipline.
The
experimental method[ology],
for ex-
ample,
is used
universally
and is
regarded
as sacrosanct within
&dquo;scientific&dquo;
psychology,
but it is
virtually
unknown within an-
thropology.
Even if
they
were
expert
with
it, however, many
anthropologists
would be offended if the
experimental
method
were
employed
on the conventional
problems
of social anthro-
pology-just
as some
experimental psychologists
are
scientifically
offended when it is not. In this
way, keepers
of the
disciplinary
faith
unintentionally
inhibit a unified science of man from de-
veloping, simply by insisting
that
anthropology, psychology,
so-
ciology,
or whatever is
properly
conducted
only
within the
bounds of the
professions customary
scientific
paradigm (see
Kuhn
1970).
For this
reason,
many
of the most
productive
ad-
vances in human
knowledge
are
yet
to
come,
and
many
will
be made
only
insofar as behavioral scientists feel free to work
at the boundaries between two or more established
disciplines.
NOTES
1This
paper
was
presented
in a somewhat modified form at the work-
shop
and
panel
on "The
Strategy
of
Comparative Inquiry," Comparative
Interdisciplinary
Studies
Section/International Studies
Association,
Annual
Meeting, Washington, D.C., February
1975. I thank Raoul
Naroll,
Robert
C.
Ness,
Evelyn
C. Rohner and Caroline C. Turner for their
helpful
com-
ments on an earlier draft.
2
Regarding
the virtual demise of
comparative anthropology,
Acker-
knecht
(1954: 117)
wrote that "within a hundred
years
the
comparative
method has in
anthropology, especially
in cultural
anthropology,
descend-
ed from a dominant
position
to a
point
where it is in
general
either not
practiced
or even condemned
explicitly." Through
the influence of men
such as Radcliffe-Brown
(1952), however,
the
comparative
method had
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
139
a more continuous utilization in
England
than in the United States (see
Nadel 1951:
222-55).
I should note that when I use the term "anthro-
pology"
in this
paper,
I mean
ethnology,
or sociocultural
anthropology,
not
archeology, linguistic anthropology,
or
biological anthropology, although
much of the
argument
is
pertinent
to these
subdisciplines,
too.
3
Comparative anthropologists
observe and measure both the commonali-
ties and the differences in behavior of
populations,
and some scholars
compare
human behavior with the behavior of
nonhumans, especially
the
higher primates.
4
When
they
are confronted with worldwide
generalizing
efforts,
how-
ever, many
of these
anthropologists
become
perceptibly
uncomfortable.
5
Sapir ( 1916 )
wrote the classic discussion of the use of the
comparative
method for
reconstructing specific
culture
history.
For an
expanded
dis-
cussion of trait distribution
studies,
see Driver
( 1970).
Driver
(1973:
171-75)
also
presented
a brief but effective discussion of the
comparative
basis of Boas "culture area school." Readers who are interested in com-
parative
trait distribution
studies,
of both the statistical and the non-
statistical
variety,
should also become familiar with the German-Austrian
kulturkreise
school,
as well as with the American and British historicalist-
diffusionist movements of the 1920s and 1930s.
Especially
notable in this
context was the
University
of California
(Berkeley)
"Culture Element
Survey" supervised by
Kroeber.
6
Holonational studies are often identified as
cross-polity surveys
or
as cross-national
surveys.
7I should
note, though,
that a collection of case studies on a
topic
is
invaluable for
comparative
research. Such case studies form the data base
for the holocultural
method,
for
example. Also,
the individual case
study
may
be used
effectively
for
exploratory research,
for
hypothesis-formu-
lating research,
and for intracultural or
within-system hypothesis testing—
but not for worldwide
hypothesis testing.
8
The
probability
model deals with the statistical
probability
of an event
occurring,
or of the co-occurrence or
sequential
occurrence of events under
specified
conditions. The mechanical model described later
postulates
more-
or-less
exceptional
uniformities or invariants of behavior. It assumes that
behavior is
ruleful,
and that if one knows the rules in sufficient detail
he can
predict
behavior with more-or-less
perfect accuracy.
9
The
history
of the
hologeistic
method
goes
back to the late
1800s,
but
Narroll
(e.g. 1973)
is
today
the
leading exponent
and
developer
of the
methodology.
For a
thorough
discussion of the
principal methodological
problems
of the method—in the context of data—see Rohner 1975b.
10
The details of the research
procedures
and tests used in the com-
munity study component
are
given
in a Field Manual
for
the
Study of
Parental
Acceptance-Rejection ( Rohner n.d.).
11
In order to measure the
parent-child relationship
and to assess the
personality
characteristics of
parents
and
children,
a
variety
of discrete
methods are used in the socialization facet of each
community study,
namely parent
and child
interviews,
a
parent-child
relations
questionnaire
administered to
parents
as well as to
sample children,
a
personality
as-
sessment
questionnaire
administered to both the
sample
children and to
their
principal caretakers,
and
systematic, time-sampled
and behavior-set-
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
140
ting-sampled
behavior observations. The field team
ideally comprises
an
anthropologist,
who concentrates
mainly
on the
ethnographic
facet of the
research,
and a
psychologically
trained
researcher,
who focuses on the
socialization facet of the research.
Many
of these instruments and
pro-
cedures are also used in the
psychological
research
component.
REFERENCES
Ackerknecht,
Erwin H.
1954 "On the
comparative
method of
anthropology,"
in Robert F.
Spencer,
ed.,
Method and
Perspective
in
Anthropology,
Min-
neapolis, University
of Minnesota Press: 117-25.
Ardrey,
Robert
1966 The territorial
imperative,
New
York,
Atheneum.
Bennett, John
W.,
and Gustav Thaiss
1967
"Survey
research and sociocultural
anthropology,"
in Charles
Y.
Glock, ed., Survey
Research in the Social
Sciences,
New
York,
Russell
Sage
Foundation: 269-311.
Berlin,
Brent,
and Paul
Kay
1969 Basic color terms: their
universality
and
evolution, Berkeley,
University
of California Press.
Boas,
Franz
1894 Indianische
Sagen
von der
Nord-Pacifischen
Küste
Amerikas,
Berlin,
Asher.
1895 "Die
Entwickelungder Mythologien
der Indianes der Nord-
Pacifischen Kiiste
Amerikas," Zeitschrift fiir Ethnologie
27:
487-523.
1896 "The limitations of the
comparative
method of
anthropology,"
Science 4: 901-08
(reprinted
in
Boas, Race, language,
and
culture,
New
York, Macmillan,
1940:
270-80).
Bowlby, John
1969 Attachment and
loss,
2
vols.,
vol.
1, "Attachment,"
New
York,
Basic Books.
Brislin,
Richard
W.,
Walter
J. Lonner,
and Robert M. Thorndike
1973 Cross-cultural research
methods,
New
York, Wiley-Interscience.
Campbell,
Donald T.
1961 "The mutual
methodological
relevance of
anthropology
and
psychology,
in F. L. K.
Hsu, ed., Psychological Anthropology,
Homewood, Ill.,
Dorsey
Press: 333-52.
Carlson,
Rae
1971 "Where is the
person
in
personality
research?"
Psychological
Bulletin 75: 203-19.
Chomsky,
Noam
1965
Aspects of
the
theory of syntax, Cambridge, Mass.,
MIT Press.
Dasen,
Pierre R.
1972 "Cross-cultural
Piagetian
research: a
summary," Journal of
Cross-Cultural
Psychology
3: 23-29.
Dawson, John
L. M.
1971
"Theory
and research in cross-cultural
psychology,"
British
Psychological Society,
Bulletin 24: 291-306.
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
141
Denenberg,
Victor H.
1969 "Animal studies of
early experiences:
some
principles
which
have
implications
for human
development,"
in
John Hill, ed.,
Minnesota
Symposium
on Child
Psychology, Minneapolis,
Uni-
versity
of Minnesota Press: 31-45.
DeVore,
Irven
1963 "Mother-infant relations in
free-ranging
baboons,"
in H. L.
Rheingold, ed.,
Maternal Behavior in
Mammals,
New
York,
John Wiley
and Sons: 305-35.
DeVos, George A.,
and Arthur A.
Hippler
1968- "Cultural
psychology: comparative
studies of human
behavior,"
69 in Gardner
Lindzey
and Elliot
Aronson, eds.,
The Handbook
of
Social
Psychology,
2d
ed.,
5
vols., Reading, Mass.,
Addison
Wesley: 4,
323-417.
Driver,
Harold E.
1961 Indians
of
North
America, Chicago, University
of
Chicago
Press.
1970 "Statistical studies of continuous
geographical distributions,"
in Raoul Narroll and Ronald
Cohen, eds.,
A Handbook
of
Method in Cultural
Anthropology,
Garden
City, N.Y.,
Natural
History
Press: 620-40
(reissued
in 1973
by
Columbia Univer-
sity Press).
1973 "Cultural
diffusion,"
in Raoul Narroll and Frada
Naroll, eds.,
Main Currents in Cultural
Anthropology,
New
York, Appleton-
Century-Crofts :
157-83.
Driver,
Harold
E.,
and William C.
Massey
1957
"Comparative
studies of North American
Indians,"
Transactions
of
the American
Philosophical Society
47: 165-456.
Fox,
Robin
1970 "The cultural
animal,"
Social Science
Information
9: 7-25.
Freedman,
Daniel G.
1968
"Personality development
in
infancy:
a
biological approach,"
in Sherwood L. Washburn and
Phillip
C.
Jay, eds., Perspec-
tives on Human
Evolution,
New
York, Holt,
Rinehart and
Winston:
1,
258-87.
Freeman,
Derek
1966 "Social
anthropology
and the scientific
study
of human be-
havior,"
Man 1: 330-42.
Freud, Sigmund
1930 Civilization and its
discontents, London, Hogarth
Press.
1950 Totem and taboo
(translated by James Strachey),
New
York,
W. W. Norton.
Goldberg,
Steven
1973 The
inevitability of patriarchy: Why
the
biological difference
between men and women
always produces
male
domination,
New
York,
William Morrow.
Greenberg, Joseph
H.
1975 "Research on
language universals,"
in Bernard
J. Siegel,
Alan
R.
Beals,
and S. A.
Tyler, eds.,
Annual Review
of Anthropology,
Stanford,
Stanford
University
Press:
4,
75-94.
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
142
Hall, G. Stanley
1904 Adolescence: its
psychology
and its relation to
physiology,
an-
thropology, sociology, sex, crime, religion
and
education,
New
York, Appleton.
Hamburg,
David A.
1968 "Evolution of emotional
responses:
evidence from recent re-
search on non-human
primates,"
Science and
Psychoanalysis
12: 39-54.
Harlow, Harry
F.
1971
Learning
to
love,
San
Francisco, Cal.,
Albion
Publishing
Com-
pany.
Jahoda,
Gustav
1970 "A cross-cultural
perspective
in
psychology,"
Advancement
of
Science 27: 57-70.
Jakobson,
Roman
1968 Child
language, aphasia
and
phonological universals,
New
York,
Humanities Press.
Keesing, Roger
M.
1973
"Paradigms
lost: the new
ethnography
and the new
linguistics,"
Southwestern
Journal of Anthropology
28: 299-332.
Köbben,
André
J.
F.
1970
"Comparativists
and
noncomparativists
in
anthropology,"
in
Raoul Naroll and Ronald
Cohen, eds.,
A Handbook
of
Method
in Cultural
Anthropology,
Garden
City, N.Y.,
Natural
History
Press: 581-96
(reissued
in 1973
by
Columbia
University Press).
Kuhn,
Thomas S.
1970 The structure
of scientific revolutions,
2d
ed., enlarged,
Chi-
cago, University
of
Chicago
Press.
Leach,
Edmund R.
1961
Rethinking anthropology, London, University
of
London,
Ath-
lone Press.
Lenneberg,
Eric H.
1964 "A
biological perspective
of
language,"
in his New Directions
in the
Study of Language, Cambridge, Mass.,
MIT Press: 65-88.
Lévi-Strauss,
Claude
1963 Structural
anthropology (translated from French
by
Claire
Jacohson
and Brooke G.
Schoepf),
New
York,
Basic Books.
Lewis,
Oscar
1966
"Comparisons
in cultural
anthropology,"
in Frank W. Moore.
ed., Readings
in Cross-Cultural
Methodology,
2d
printing, reset,
New
Haven,
HRAF Press: 50-85.
Lorenz,
Konrad
1966 On
aggression,
New
York, Harcourt,
Brace,
& World.
Malinowski,
Bronislaw
1927 Sex and
repression
in
savage society,
New
York,
World.
Mead, Margaret
1928
Coming of age
in
Samoa,
New
York,
William Morrow.
Morris,
Desmond
1967 The naked
ape:
a
zoologists study of
the
human animal,
New
York,
McGraw-Hill.
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
143
Murphy,
Robert,
and
Julian
H. Steward
1956
"Tappers
and
trappers: parallel processes
in
acculturation,"
Economic
Development
and Cultural
Change
4: 335-53.
Nadel, Siegfried
F.
1951 The
foundations of
social
anthropology, London,
Cohen and
West.
Naroll,
Raoul
1968 "Some
thoughts
on
comparative
method in cultural
anthropol-
ogy,"
in Hubert M.
Blalock, Jr.,
and Ann B.
Blalock, eds.,
Methodology
in Social
Research,
New
York,
McGraw-Hill:
236-77.
1973 "Holocultural
theory tests,"
in Raoul Naroll and Frada
Naroll,
eds.,
Main Currents in Cultural
Anthropology,
New
York, Ap-
pleton-Century-Crofts:
309-84.
Naroll, Raoul,
Timothy J. OLeary,
and Lee
Sigelman
1974 "A
hologeistic bibliography," Working Paper
No. 23
ISA/CISS,
Pittsburgh, University
Center for International Studies.
Osgood,
Charles
E.,
William H.
May,
and
Murray
S. Miron
1975 Cross-cultural universals
of affective meaning, Urbana,
Uni-
versity
of Illinois Press.
Piaget, Jean
1970 Genetic
epistemology (Eleanor Bucksworth,
trans.),
New
York,
Columbia
University
Press.
Radcliffe-Brown,
Alfred R.
1952 Structure and
function
in
primitive society, Glencoe, Ill.,
Free
Press.
Redfield,
Robert
1941 Folk culture
of Yucatan, Chicago, University
of
Chicago
Press.
Rohner,
Ronald P.
1970 "Parental
rejection,
food
deprivation,
and
personality develop-
ment : tests of alternative
hypotheses," Ethnology
9: 414-27.
1975a "Parental
acceptance-rejection
and
personality development;
a
universalist
approach
to behavioral
science,"
in W. W.
Brislin,
S.
Bochner,
and W.
J. Lonner, eds.,
Cross-Cultural
Perspectives
on
Learning, Beverly
Hills,
Calif., Sage
Publications: 251-69.
1975b
They
love
me, they
love me not: a worldwide
study of
the
effects of parental acceptance
and
rejection,
New
Haven,
HRAF Press.
1976 "Sex differences in
aggression: phylogenetic
and enculturation
perspective,"
Ethos 4: 57-72.
n.d. Field manual
for
the
study of parental acceptance-rejection
(in preparation).
Rohner,
Ronald
P.,
Billie R.
DeWalt,
and Robert C. Ness
1973
"Ethnographer
bias in cross-cultural research: an
empirical
study,"
Behavior Science Notes 8: 275-317.
Rohner,
Ronald
P.,
and Leonard Katz
1970
"Testing
for
validity
and
reliability
in cross-cultural
research,"
American
Anthropologist
72: 1068-73.
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from
144
Sapir,
Edward
1916 "Time
perspective
in
aboriginal
American culture: a
study
in
method," Ottawa,
Government
Printing Bureau, Geological
Survey (Memoir 90):
1-87.
Sears,
Robert R.
1961 "Transcultural variables and
conceptual equivalence,"
in Bert
Kaplan, ed., Studying Personality Cross-Culturally, Evanston,
Ill., Row,
Peterson: 445-55.
Sebeok,
Thomas A.
1968 "Goals and limitations in the
study
of animal
communications,"
in his Animal Communication
Bloomington,
Indiana Univer-
sity
Press: 3-14.
Tiger,
Lionel
1969 Men in
groups,
New
York,
Random House.
at ZB MED on September 15, 2014 ccr.sagepub.com Downloaded from

You might also like