Memes and their themata

Author(s): WILLARD L. MIRANKER

Source: The American Journal of Psychology, Vol. 123, No. 3 (Fall 2010), pp. 307-317
Published by: University of Illinois Press
Stable URL: http://www.jstor.org/stable/10.5406/amerjpsyc.123.3.0307
Accessed: 02-07-2016 04:24 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.

University of Illinois Press is collaborating with JSTOR to digitize, preserve and extend access to The
American Journal of Psychology

This content downloaded from 70.95.156.2 on Sat, 02 Jul 2016 04:24:56 UTC
All use subject to http://about.jstor.org/terms
 Memes and their themata
WILLARD L. MIRANKER
Yale University

When it is instantiated as a neuronal state, a meme is characterized as a phenotype in a novel

neuronal sense. A thema is an instantiation of a meme as a conscious experience (a thought-
meme). It is a primitive to which no location may be attributed, and it serves as a canonical
representative of a class of memes. Memes in such a class may have physical or ideal (Platonic)
instantiations. Pairing of this memetic phenotype characterization with the ideal thematic primi-
tive is an example of other pairings in nature that are identified, and in particular it informs a de-
scription of the pairing of the unconscious mind and manifestations of consciousness. Interrela-
tionship of these pairings is what illuminates aspects of each of them. These constructs support
introduction of a consciousness thesis and then a notion of a dynamic self-referential grammar
that generates a growing repertoire of consciousness manifestations. A method showing how
a neuronal state generates a specific concept (thema) is introduced, and a sample of a class of
examples is given. Pointers to experiments relevant to development of the thesis are given.

Biotic constructs (such as phenotypes1 and geno-
types) and developmental processes (such as com-
petition, selection, and reproduction) are aspects of
natural mechanisms of which customary Darwinian
evolution (Darwin, 1859) is but one example do-
main. I propose that memes, neuronal arrays, and
mind (conscious or unconscious) each furnish other
such domains. Understanding of those domains flows
from analogies between them. Analogizing constitu-
ents of a scientific domain with those of a cultural
one generates conceptual fertility. For example, note
that the development of memetics (whose roots are
found in Semon, 1921) flows from Dawkins (1999),
who analogizes between genetic-based biotic evolu-
tion and culturally based evolution of concepts in
just such a way. Our use of an analogy between the
pairings within both memetic and genetic constructs
has a forerunner in the work of Lumsden and Wilson
(1981) and of Wilson (1999), in which a correspon-
dence between fundamental units of culture and of
neural networks is discussed.
I begin by describing a correspondence between
the pairing of synaptic functioning with neuronal
functioning and the pairing characterized by the
customary biotic genotype–phenotype relationship.
An extension is made to an analogous pairing cor-
respondence within memetics, namely to the pairing
of memes and their associated thema. A meme is an

American Journal of Psychology

Fall 2010, Vol. 123, No. 3 pp. 307–317 • © 2010 by the Board of Trustees of the University of Illinois

This content downloaded from 70.95.156.2 on Sat, 02 Jul 2016 04:24:56 UTC
All use subject to http://about.jstor.org/terms
 idea or concept as well as any instantiation of that
idea or concept. See Dawkins (1999) and Blackmore
(1999), where the role of culture–gene co-evolution, a
critical feature of memetics, is elaborated. The notion
of a thema, an instantiation of a meme as a conscious
experience, that is, as a thought (a thought-meme),
will be introduced. As shall be seen, the character-
ization of the thema as an aspect of consciousness
augments the role of the memetic–thematic pairing
in the culture–gene co-evolution. Consciousness is a
primitive and thus an undefined concept. A primitive
is typically specified by its properties. In the case of
consciousness, take the following features as an in-
formal working definition: the perception of qualia
(note that the term quale denotes a property as it is
experienced as distinct from any source it might have
in a physical object, e.g., a color, a sound, a pain), the
manifestation of a feeling (e.g., hunger, fear, love), a
notion of intentionality (a sense of being in charge
of one’s actions), and a notion of self (cognitive and
affective adaptation of one’s identity). (See the issue
on the self in the Journal of Consciousness Studies,
Volume 5, Issue 2, 1998.)
A meme’s customary characterization is that of a
selfish replicator, spreading culturally from brain to
brain, this being an extension of Dawkins’s (1999)
notion of the gene-centered view of evolution. I start
with a modification of this customary replicator char-
acterization by framing the meme as a phenotype
when it is instantiated as a neuronal state, that is, the
state of a neuronal circuit in the brain at an instant of
time. Such a memetic instantiation may be real, that
is, a literal neuronal state, or it may be ideal (Platonic),
possibly corresponding to an abstract model of that
state. These constructs inform a description that I
give of the unconscious mind and manifestations of
consciousness as a corresponding pairing between
a neuronal state and what I call a thema. The thema
serves as a canonical representative of the memes
belonging to a specific class (i.e., all the forms of a
particular concept), and so it can be viewed as the
subject, the literal theme of that class. Although the
thema represents the concept instantiated by a cor-
responding class of memes, it is given the novel role
of a conscious experience in this development. Se-
mon (1921) introduced the concept of a unit of social
evolution that he called a mneme. Although there is
308  •  miranker
This content downloaded from 70.95.156.2 on Sat, 02 Jul 2016 04:24:56 UTC
All use subject to http://about.jstor.org/terms
no agreement on what makes up one unit of cultural
transmission, the thema can be viewed as a candidate.
These proposed constructs support the introduc-
tion of a consciousness thesis. If consciousness is an
ongoing, learned capability of the unconscious mind,
a notion of a dynamic self-referential grammar for
consciousness on which that learning supervenes can
be introduced. Other treatments of a theory of mind
that use memes are found in Blackmore (1999) and
Sheehan (2006).
I propose a method for constructing themata as
mathematical abstractions by using a model neural
net and results of axiomatic set theory (Moschko-
vakis, 2005; Miranker & Zuckerman, 2009). Arbi-
trarily many variants of such a construction are pos-
sible, and a particular meme may correspond to any
of the corresponding neuronal states. An example,
a neuronal structure corresponding to the thema
of the class of memes that correspond to the con-
cept of a particular natural number, is developed.
More elaborate examples are found in Miranker and
Zuckerman (2010). This connection between neu-
ronal structure and themata resembles the studies of
neural correlates of consciousness as found in Koch
and Crick (2001) and in Koch (2004). A neural cor-
relate of a conscious perception may be taken as the
neural assembly in the brain that is activated when
awareness of that perception occurs. Later in this
article I develop a framework for the deeper study
of consciousness and its neural correlates.
OUTLINE
First, I describe neuronal firing to be the result of
a competition with selection between neurons, en-
abling characterization of the pairing of synaptic
function with neuronal functioning as a genotype–
phenotype correspondence. Compare this with the
neural Darwinism of Edelman (1978). Genetic repli-
cation is interpreted as a form of strengthening, and
so it is the strengthening of a synapse (the increasing
of the synaptic weight) that is taken as the analog of
gene replication.
Next, I observe that a meme instantiated as an
input to a neuronal assembly generates a recording
of information in the afferent synaptic weights. Such
information representations inform a competition
between neurons for firing, and so they also inform
a process of selection among afferents for differen-
tial strengthening (Hebb, 1949). For this reason the
meme’s neuronal representation array is viewed as a
memetic analog of a phenotype. To the relevant array
of afferent synaptic weights, the role of the memetic
genome is assigned, and so in analogy to the gene, to
this array’s recorded information (a meme) the status
of a replicator is attributed.
Themata are introduced to show that the pairing
of the two mind agencies of consciousness and the
unconscious supplies a favorable quality for evolu-
tion. I propose existence of a dynamic (expanding)
self-referential grammar that frames development of
a growing consciousness repertoire, consisting of the
collection of formal sentences (memes) that can be
generated or parsed through that grammar.
In the next section the study of memes is for-
malized, through the mathematical construct of an
accessible pointed graph and its labeled decoration
(Aczel, 1988). Then I frame an example of a class of
memes (with both physical and Platonic members)
and its Platonic thema.
In the concluding remarks I point toward experi-
mental approaches for the study of the constructs
presented.
THE SELFISH SYNAPSE
Synapse strengthening as an information conveyance
analogy to a gene replicating
Both genes and synapses have the role of informa-
tion-conveying agents. Whereas the gene has been
described as a replicator, a selfish replicator, the syn-
apse has not. However, both genetic replication and
synaptic strengthening are processes that increase the
effectiveness of an information-conveying agent, the
gene or synapse, as the case may be. The stronger the
synapse, that is, the greater its synaptic weight, the
more effective it is. Genes do not strengthen; rather,
they express their effectiveness through replication.
When synapses are joined in a cohort, increasing ef-
fectiveness of the cohort is accomplished cooperative-
ly, possibly by strengthening of some of its members
while others are weakened. So I view genetic replica-
tion and synaptic strengthening as special cases of
increasing the effectiveness of information-conveying
agents in nature. There are corresponding artificial
This content downloaded from 70.95.156.2 on Sat, 02 Jul 2016 04:24:56 UTC
All use subject to http://about.jstor.org/terms
agents, an example of which arises in the study of
genetic algorithms (Fogel, 2000).
A neuron firing corresponds to a phenotype succeeding
A gene has fulfilled its role if the phenotype it is a
part of succeeds at competition for survival and re-
produces, thereby replicating the gene (conveying
information). A synapse fulfills its role if the neuron
of which it is an active afferent fires (conveying in-
formation). In this view, the neuron takes the role of
a phenotype. It is deemed successful in a competi-
tion with other neurons in its environment (the latter
specified by neuronal connectivity and activity) for
being able to fire if and when it does fire. So I view
biotic reproduction and neuronal firing as special
cases of success in competition for survival (and so
informing a selection process).
Selection
According to Hebb’s rule, an afferent synapse in-
creases or decreases in strength if that synapse’s ac-
tivation is positively or negatively correlated with its
neuron’s firing (Hebb, 1949; Haykin, 2008). This
characterizes a selection process. So if an afferent
synapse correlates with the neuron’s firing, the syn-
apse has fulfilled its role in contributing to the pro-
duction of a successful phenotype. Such synapses
are selected from among all the neuron’s afferents
to be strengthened. An afferent that is negatively
correlated with the firing decreases in strength. Re-
peated decreases may lead to the elimination of that
synapse. Sustained quietude of a neuron (an expres-
sion of persistent lack of success in competing to
fire) can likewise lead to that neuron’s elimination
(apoptosis). I view this Hebbian strengthening and
weakening protocol of a synapse as part of a neu-
ronal process of natural selection.
Propagation of information
I characterize these observations by saying that suc-
cessful synapses propagate in the sense that they
strengthen and endure as information conveyers.
The synapses that propagate increase their influ-
ence on the neuron’s functioning in the sense that
the neuronal inputs and the array of afferents, count-
ing synaptic strength (these features constituting the
neuronal analog of the environment of a customary
memes and their themata  •  309
 biotic phenotype), inform neuronal firing. That is, the
stronger a synapse is, the greater its ability to influ-
ence firing (Haykin, 2008).
Recording and processing of information in a
neuronal assembly
Genes and groupings thereof represent the recording
of information at different scales. (In malarial infections
hundreds of genes participate in such a grouping; Su
& Wellems, 1996.) Analogously, I view the synaptic–
neuronal interaction as a process at increasing scales:
synapse–neuron–neuronal assemblies of increasing
size. The strengthening or weakening of synapses in-
forms a process of recording of information at these
different scales. Associative memories stored in the
brain are recorded in just this way. Propagation of in-
formation by a neuronal assembly is a processing of
neuronal activity flowing through the assembly, which
is informed by the neuronal environment.
MEMES
The meme as phenotype
As described in the preceding section, a meme may
be instantiated as a distributed input encoded and
stored in a neuronal assembly, with that encoding
embodying a recording of information, which repre-
sents that meme at one or another scale depending
on the meme’s complexity. This information array,
the recording in the neuronal assembly, specifically
an encoding in terms of the synaptic weights, is a
distributed representation of the meme. As described
earlier, such representations play a key role in the
competition between neurons for firing, and these
features inform a selection process among synapses
for strengthening or weakening, as the case may be.
Therefore, I take the view that the memetic neuronal
representation array may be viewed as a memetic ana-
log of a phenotype. The synaptic weights in this array
(this phenotype) can be viewed as a memetic analog of
a genome, because just as the information recorded in
the biotic genome is part of and informs specification
for the biotic phenotype, the information stored in the
synaptic component (via the weights) of the neuronal
array (or what I call the memetic genome) both en-
codes and can be exploited for reproduction, that is,
for retrieving information in order to copy the meme.
So I view this memetic genome as a replicator. The
310  •  miranker
This content downloaded from 70.95.156.2 on Sat, 02 Jul 2016 04:24:56 UTC
All use subject to http://about.jstor.org/terms
copy, also a phenotype although unlike the conven-
tional case, may take many different forms (memetic
instantiations), such as another neuronal construct,
an utterance of speech, or a sample of writing.
Jumping memes
A speech utterance, for example, provides the input
to a second person and so, via the by now familiar
Hebbian mechanism, enables the meme to jump and
be recorded by the second person. There is the cus-
tomary blurring of the genotype–phenotype relation-
ship in the meme–neuronal pairing. That is, there is a
sharing of the genotype and phenotype roles. Indeed,
although the biotic genotype and phenotype have
complementary functionality, recall that the former
is a material part of the latter. Competition and sur-
vival for memes are not specified as they are in the
customary biotic context. This is necessitated in part
by a difference of both temporal and spatial scales at
which the respective associated dynamics (replica-
tion and instantiation) develop. I stress that it is the
absence of overly narrow constructed equivalences of
the analogs between pairings that illuminates under-
standing of phenomena within those pairings.
Storage and retrieval of memetic information
The distributed nature of synaptic storage of me-
metic information constitutes an associative mem-
ory (Haykin, 2008). This makes for a richer context
than the DNA-based one found in customary biotic
genome–phenotype arrangement; the latter may be
likened to an addressable memory (usually but not
accurately called a random access memory). The
neuronal assembly supporting a memetic genotype–
phenotype arrangement can be multiplexed to record
other inputs, that is, to record other memes. So a num-
ber of different memes may share a single neuronal
array, that is, they may share a memetic genome and
phenotype. Although genes may be viewed as fixed,2
their memetic counterparts, the synaptic strengths,
are in flux. Although it is not known how the biotic
phenotype is constructed given its genome, the meme
may be recovered. That is, the meme may be read out
from the memetic phenotype, more explicitly, from
the memetic genome encoded by means of synaptic
weights in the relevant neuronal information array. As
a readout from an associative memory, this process
requires that an approximation to the encoding of
the meme be used as an input cue to that array. Is
it an epigenetic correspondent to this neuronal cu-
ing that is involved in activating or deactivating the
genes, thereby supplying the information that drives
embryonic development?
The memetic replicator
The meme is customarily described as sharing with
the gene the characteristic of being a selfish replicator.
The prior section shows that the memetic replicator
is reified in the associative memory, explicitly in the
collection of synaptic weights that encodes the meme.
In this picture, all of the possible memetic instantia-
tions, one of which is the relevant neuronal array, are
phenotypes and not replicators. As suggested earlier,
I assign to the memetic genome (i.e., the relevant as-
sociated array of synaptic weights) the status of a
replicator in analogy to the genetic case.
SCALES, MIND, THEMATA, CONSCIOUSNESS
­THESIS, AND GRAMMAR FOR CONSCIOUSNESS
Scales in evolution, genomic–phenotypic pairing
Biotic matter first appeared 4 billion years ago at the
molecular scale with simple proteins. Some consider
crystalline structures and prions to be primitive or
proto biotic forms (Belkin, 2003). With evolution, the
complexity in the biotic organization increases until
the simplest form with a partitioning into a genome–
phenotype pairing is arrived at. This pairing could
be identified with the appearance of eukaryotic cells
(cells with a nucleus that separates the metabolic
and genomic structures) 2 billion years ago. (Sym-
biosis among prokaryotes is viewed as critical in the
formation of eukaryotes; Margulis, 1991.) Might the
constituents of our memetic pairings be viewed as
symbionts? Appealing to Darwinian principles, I
can conclude that the separation and specialization
of roles characterized by this pairing appeared and
endured through competition and selection as a fa-
vorable arrangement for survival and propagation.
Memes may be admitted to the picture when the biotic
complexity increases to the level of neurons (the me-
metic phenotypes) and their afferent synaptic arrays
(the memetic genomes), as has been asserted. The
pairing of a neuron and its afferent synapses also con-
stitutes an evolved partitioning of roles (phenotype
and genome or, alternatively, information processing
This content downloaded from 70.95.156.2 on Sat, 02 Jul 2016 04:24:56 UTC
All use subject to http://about.jstor.org/terms
and instantiation of that information) that is probably
favorable to the selection process of evolution. Did
memes appear in primitive neuronal organizations
or only when assemblies of an adequately elaborated
such pairing structure developed? Could memes ap-
pear before neurons (consider the cognition exhib-
ited by bacterial colonies; Ben Jacob, Becker, Shapira,
& Levine, 2004)?
Mind and its consciousness manifestations, other pairings
The mind could be taken to arise from the work-
ings (layers of information processing at appropriate
scales) within the brain. For reasons of clarity I start
with the conventional view that mind supervenes on
brain. However, it is the entire physical body and
in some sense the environment as well (some take
an even grander frame; Chalmers, 1996) that is in
play.3 These mind workings are the processing and
movement of information in some neuronal assembly.
As this assembly increases in complexity, so may the
relevant workings. This is a picture of operation at
various scales. So is a degree of complexity in the
unconscious workings in brain–mind needed for the
emergence of the manifestations associated with con-
sciousness? At the current state of our knowledge,
consciousness must be taken as a primitive (Miranker
& Zuckerman, 2009). As such it has no formal defini-
tion but must be specified informally by means of its
behavior and properties. Examples of such manifes-
tations are a notion of self (Journal of Consciousness
Studies, 1998), intentionality, and qualia. Examples
of qualia are vision (three-dimensional space with
its illusion of depth, colors, textures, and shadings),
touch (e.g., pain, wetness, cold), and feelings (e.g.,
fear, joy, love, hope). I do not rule out the possibility
that consciousness in some form could arise with the
organization of the first biotic matter, that is, before
the appearance of brains. See Ben Jacob et al. (2004),
where descriptions of bacterial colonies that exhibit
communication and cognitive functionality are found.
In Miranker (2010), the modeling and simulation
studies of the self, qualia, and intentionality in bac-
terial colonies is found. See also Sheets-Johnstone
(1998) for examples and philosophical commentary
on this issue.
Is a favorable quality obtained for evolution (for
selection, strengthening, and replication) by such a
separation and specialization of function of the two
memes and their themata  •  311
 mind agencies of consciousness and the unconscious?
(Consciousness and the unconscious each probably
represent layers of differentiated agencies of mind.)
For example, are the manifestations of consciousness
a representation of information in a form to make the
storage (encoding) and processing of that information
more efficient or more robust (e.g., more stable with
respect to disturbances)? Can the pairing of these
two mind agencies be described as another variant of
the genotype–phenotype relationship that I am pro-
pounding? If so, what is the associated competitive
selection advantage for the underlying agents (pheno-
types and their associated genotypes)? In particular,
what are the replicators, and what is their process of
strengthening? Will evolution go beyond conscious-
ness and the unconscious with this paradigm of pair-
ings? To start to address these questions, I turn to the
notion of themata.
Themata, competition, a winning thema,
a consciousness thesis
Consider the information being processed in a neu-
ronal array that was referred to earlier as a component
of the mind. The term thema (thought-meme) is used
to describe this information as a virtual construct, that
is, as a Platonic and not a physical object. A thema
might be viewed as a kind of virtual photographic
negative of a meme, and the meme’s associated instan-
tiating neuronal array, a phenotype, may be viewed as
a kind of print of that negative. An exogenous stimu-
lus supplies the input needed to initiate a Hebbian
synaptic altering process of the synapses of the neu-
ronal array (phenotype) that instantiates (stores) the
meme on which the Platonic thema supervenes. This
associative storage of information is multiplexed, and
so the strengthening of one memory record, that is, of
one meme, may degrade others. This type of trade-off
is present at the lower synaptic scale also, where in a
collection of synapses, such as the memetic-thematic
genome corresponding to the meme being stored,
some synapses increase in strength during memetic
encoding or recording while the strength of others
might decrease.
Themata compete to be strengthened, which is
their way of propagating. (Recall that competition
between neurons and their processing is a central
aspect of their study; Edelman, 1978; Miranker, 2005;
Haykin, 2008.) Suppose that there is a winning the-
312  •  miranker
This content downloaded from 70.95.156.2 on Sat, 02 Jul 2016 04:24:56 UTC
All use subject to http://about.jstor.org/terms
ma. For definiteness, say the one that corresponds to
the array with the most firing neurons, this being one
measure of the currently most active or most success-
ful memetic phenotype. Taking consciousness to be
an ideal quality (a Platonic primitive4), I propose the
following consciousness thesis based on this hypoth-
esized competition between themata.
Consciousness thesis: The winning thema
generates a manifestation of consciousness rep-
resentative of that thema.
This thesis could be viewed as a formalized
extension of the Pandemonium model of Selfridge
(1959). This model consists of a multiplicity of paral-
lel demons, each specialized in recognizing a specific
visual stimulus. Recognizing his stimulus, a demon
calls out to the master demon. This exclamation is
louder the more highly the demon estimates the
probability of a correct recognition (identification).
Selfridge’s demons could be viewed as precursors of
the themata, with the loudest-shouting demon being
the winner.
Consciousness as an agency for propagating the mind,
neural correlates
Generation of a manifestation of consciousness re-
quires an appropriate cue, an exogenous sensory in-
put (as do the retrievals from an associative memory).
So the mind supervenes not only on the entire physi-
cal body but on the environment as well (Chalmers,
1996), the latter being the usual source of the cue. If
this cue produces a winning thema, then according
to the consciousness thesis, the winner’s phenotype,
a neuronal array that shall be called this thema’s neu-
ral correlate generates a representative manifestation
of consciousness. Because of this manifestation’s
form as an awareness of an aspect of the environ-
ment, I propose that the neural correlate takes this
secondary awareness to be a supplementary sensory
input. (Release hallucinations furnish examples of
reentrant perceptual-like processes; Sacks, 2007.) I
suppose that this putative input cues and reinitiates
the neuronal processing in question and so further
strengthens that thema through its memetic genome
(the relevant synapses) via Hebbian processes. That
is, the underlying meme is further strengthened
(propagated). In this scenario, as an evolutionary
development,
 A function of consciousness is to act as an
agency for strengthening and enlarging the un-
conscious, that is, as an agency for propagating
the mind.
Grammatical thesis for consciousness, observations
on consciousness manifestations
The development motivates introduction of the fol-
lowing grammatical thesis (cf. Chomsky, 1953).
Grammatical thesis for consciousness: There
is an intrinsic dynamically increasing self-refer-
ential grammar5 that supports the development
of a consciousness repertoire of interpretable
manifestations of consciousness.
I claim that creating a manifestation of conscious-
ness (through the cued memetic retrieval from neu-
ral correlates described earlier) is a learned ability
informed by this grammar. This ability is used to
develop a consciousness repertoire. The repertoire
grows and consolidates with use. (Recall that the hip-
pocampus is the critical brain structure implicated in
memory consolidation; Kairiss & Miranker, 1998.)
The creating ability itself improves with repetition.
Referring back to the negative and print metaphor, the
prints (memes) that are the constituents of the reper-
toire are stored multiplexed within various neuronal
arrays as associative memories. The corresponding
themata, being Platonic primitives, have no attribut-
able location. Even adults with a developed conscious-
ness manifestation–creating ability and a correspond-
ing repertoire can be presented with a sensory cue
that is not readily, indeed not at all, resolvable. That
is, the information in such a cue defies being parsed
by their grammars into an understandable or logical
concept. Indeed, the cue is not a legal sentence of
the consciousness grammar. Such a grammar consoli-
dates, becoming more robust and more efficient with
repetition and learning. It can be used to parse an in-
creasing range of concepts (sentences of the grammar).
That is, the grammar increases its effectiveness as an
augmenter of the consciousness repertoire. I project
that the learning that underlies the dynamic aspects of
the grammar is an example of self-reference analogous
to the ability of a stored program in a contemporary
computer to operate on itself. Self-reference is the
key logical feature of foundations of consciousness
described in Miranker and Zuckerman (2009).
This content downloaded from 70.95.156.2 on Sat, 02 Jul 2016 04:24:56 UTC
All use subject to http://about.jstor.org/terms
The patient Virgil supplies an illustration of these
dynamic features (Sacks, 1995). He was blinded as
a child, and a surgery restored Virgil’s sight when
he was in his 50s. At first, he was unable to fathom
(parse, in our terminology) the cues (light flashes and
colors) presented to his newly functioning visual sys-
tem. He developed some of the requisite abilities and
a related repertoire of manifestations (sentences of the
grammar) of visual consciousness but never to the
degree of a normally developed sighted individual.
He never became comfortable with his reacquired
visual ability, preferring to sit in the dark using Braille
and preferring to rely on his cane to get about. It is
likely that Virgil was unable to fully recover vision
because he was well past the critical period of neural
plasticity in which visual systems fully develop.
However, a loss of neuronal plasticity in terms
of the creation of synaptic connections does not
limit the learning, memory, and behavioral abilities
of the brain, because its neurons retain the plasticity
of synaptic development, that is, of synaptic weight
change characterized by Hebb’s law, for instance
(Hebb, 1946, 1949; Haykin, 2008). There is appro-
priate plasticity in the developed adult brain for the
formation of new memes and themata, a feature that
most adults can attest to.
Memes and their themata, including the compe-
tition in which they engage, comprising an evolving
framework (which has been characterized as function-
ing through an expanding self-referential grammar), is
used in the development of the repertoire of manifes-
tations of consciousness. Natural selection pressures
influence the development of such a repertoire and
the learning of how to manipulate it. The rare cases of
feral children demonstrate some of the subtleties and
limitations of this process. Language reveals another
aspect. Some speakers cannot hear certain sounds
used in a foreign language. The memes (“sentences”)
corresponding to such sounds cannot be parsed.
For example, the English l is unknown in Russian,
which possesses two distinct l-sounds that the normal
English speaker would find difficult to reproduce: a
hollow, guttural-like l and a soft, palatalized l-sound
that is only very approximately rendered, in English
terms, as ly. See Sapir (1921) for additional examples.
How the manifestations of consciousness in the rep-
ertoire take the forms of the qualia so familiar to us
remains an open question.
memes and their themata  •  313

A REPRESENTATIVE FOR A THEMA, THE CONCEPT
OF NATURAL NUMBERS, GENETICS
I use set theory to formalize the study of memes
and themata. A given meme may have many forms
(phenotypes), physical and Platonic. The thema, be-
ing Platonic, could have a virtual representative. In
this section I present a method for generating such
a representative and develop an example. Although
this example corresponds to one of the fundamental
concepts of thought, the natural numbers, it has a
particularly simple form. To Kronecker is owed the
remark, “God created the integers, all else is the work
of man.”
Themata and accessible pointed graphs
An accessible pointed graph (APG) is a construct con-
sisting of two classes of primitives: nodes and directed
edges between certain pairs of those nodes. There is
a distinguished node called the point of the APG. Ac-
cessibility means that every node in the graph is reach-
able from the point by a connected path composed of
directed edges. (See Figure 1b for an example of such
an APG.) Each instantiation of a meme can be made
to correspond to an APG, the graph depicting a kind
of parsing of that meme. A node is called the parent
of its children, namely of the other nodes in the APG
that are linked to the parent by a directed edge from the
parent. So the children of a node correspond to memes
on which the parent’s meme supervenes. Leaves of the
APG are the nodes that have no children. So a leaf
might be considered to correspond to a fundamental
meme. Decorating an APG is a formal process that
associates a unique class of sets (the decoration) to the
APG, one set of the decoration with each node. Call θ
the set in the decoration that corresponds to the point.
The decorated APG is called a picture (because it is a
kind of representation) of the set θ. See Aczel (1988)
for details concerning these constructs, among which
is the assertion that the APG in Figure 1b is a particular
representation of the concept of the Von Neumann
ordinal 2. Although to each APG there corresponds a
unique decoration, there are in general infinitely many
different APGs with a common set θ decorating their
point. So these possibly infinitely many APGs will
represent a population of different phenotypes of a
common meme. θ, a Platonic representative of these
314  •  miranker
This content downloaded from 70.95.156.2 on Sat, 02 Jul 2016 04:24:56 UTC
All use subject to http://about.jstor.org/terms
memetic phenotypes, is the thema sought. In fact, all
these constructs are Platonic, but as shall be seen, some
decorated APGs are models of a neural state, that is,
they are examples of a model of a neural network at a
moment of time. The neural state being modeled is
a corresponding physical instantiation of the meme
in question. Memetic phenotypes may correspond to
physical instantiations other than those provided by
a neural state.
Example
Consider the model neural network composed of
three McCulloch–Pitts model neurons, a, b, and c,
with the synaptic weights wba, wca, and wbc shown
in Figure 1a (Haykin, 2008). Neuron a is firing, but
neurons b and c are quiescent. That is, the neurons’
output activities are v(a) = 1 and v(b) = v(c) = 0. With
these data specifications, the network becomes what
has been called a neural state. The decorated APG
corresponding to this neural state, the graph shown
in Figure 1b, is a picture of a set θ. (The nodes of this
graph are labeled by the sets θ [the point], B, and
C.) In particular, θ = {B,C}, where the set B = 0⁄ (the
empty set) and the set C = {0⁄ } (the set whose only ele-
ment is 0⁄ ). {B,C} denotes the set with the elements B
and C. Of course, this example is directly extendable
to the concept of any natural number.
Memes and their thema, neural correlates of consciousness
The diagram in Figure 1b, illustrating a decorated
APG, arises from the neural state shown in Figure 1a.
Because this APG is a representation of the Von Neu-
mann ordinal 2, it is an instantiation of the concept
(the meme) of the integer 2. The set θ decorating the
point is the thema of this meme. The thema θ and
the diagram in Figure 1b are Platonic instantiations.
The corresponding actual neural state being mod-
eled (such as by the model in Figure 1a), also a meme,
is a physical instantiation of that meme. Then such
a physical neural state is a player in the competition
for generating the manifestation of consciousness of
the concept of the integer 2. I shall say that a winning
neural state with thema θ is the latter’s neural cor-
relate. Then the neural state schematized in Figure 1a
is a neural correlate of the concept (of the meme) of
the integer 2. (The possibility of a cohort of winners
that might result in an extensive neural correlate is

Figure 1a . Neural state with neurons a, b, and c
Figure 1b. Corresponding APG with point θ and leaf B
not precluded.) Recall that a single thema, such as
the example θ, may correspond to any element of an
infinite collection of neural circuits (memes). This
mathematical development adds an analytic frame-
work absent from discussions of neural correlates of
consciousness usually found in the literature (Koch
& Crick, 2001; Koch, 2004). The neural state (e.g.,
the one in Figure 1a) and the APG (e.g., the one
in Figure 1b) are examples of a potentially infinite
number of neuronal assemblies and corresponding
APGs that can have this same thema θ. So by analogy,
each meme in the class whose thema is θ is a picture
of that θ. In terms of the photographic negative and
print metaphor discussed earlier, I view such a pic-
ture as a print corresponding to its photographic
negative θ.
Memes and genetics
A connection between genetics and differentiation
in embryonic development to memes and themata
is now explored.
This content downloaded from 70.95.156.2 on Sat, 02 Jul 2016 04:24:56 UTC
All use subject to http://about.jstor.org/terms
Memes and chromosomes.
Take a chromosome to correspond to (to be mod-
eled by) the point of an APG. Take the point’s chil-
dren to be the chromosome’s primary constituent
groupings of genes. This arrangement along with the
decoration of the APG generates a set–element rela-
tionship. Such groupings, which have a temporally
changing character, are specified by the biology. To
layers of such genetic groupings, there correspond
layers of parent–child relationships in the APG.
Moreover, genes or a grouping thereof have a time
varying state of activity and inactivity. Other genes
and epigenetic processes turn them on and off. This
generates dynamics defined on the associated APGs,
because genes or groupings thereof, corresponding to
nodes in the APG, could be taken to be present only
when that gene or grouping is active. This induces
dynamics on the associated decorations (Miranker
& Zuckerman, 2010). These dynamics may cause an
associated appearance or disappearance of loops of
edges inside such APGs, that is, a selected flipping be-
tween the associated sets’ types (possibly a flipping of
these sets between being what are called well-founded
and non–well-founded sets; Aczel, 1988.) Compare
these loops with Sheehans’s looping pathways, by
means of which he asserts that patterns can iteratively
propagate, mutate, and evolve. For Sheehan (2006),
these pathways collectively work toward creation of
an intelligent mind. The appearance of these two set
types suggests a connection of the present framework
with the mathematical foundations of consciousness
of Miranker and Zuckerman (2009). The neural as-
semblies that may form a candidate neural correlate
of a meme generate in turn a class of decorated APGs,
each of which serves as an ideal memetic phenotype
(Miranker & Zuckerman, 2010). As noted earlier,
the set θ in the decoration that corresponds to the
point of such an APG is a Platonic representative of
the thema that corresponds to the class of memes in
question. This set θ is the thema itself.
Differentiation in development.
The messenger RNA and the proteins encoded by
the genes (or groupings thereof) are also characteriz-
able within this framework of pairings. Such proteins,
constituting a first stage in the production of the as-
sociated biotic phenotype, correspond to a leaf in
an APG, the latter graph characterizing a hierarchy
memes and their themata  •  315
 in the differentiated embryonic development of the
phenotype that ensues.
CONCLUSION AND EXPERIMENTS
Concluding comment
The Darwinian paradigm of natural selection and
its subsequent development in terms of the pairing
of genotypes and phenotypes has been extended to
other kinds of pairings found in nature. The first such
is to the propagation of cultural information by pair-
ing the concept of a meme and what I call a thema.
This pairing is reified in the structure of a neuronal
array and the array’s associated synaptic weights.
This neuronal framework supplies the constructs of
genotypes and phenotypes for the meme–thema pair-
ing. The neuronal development motivates a further
extension of the Darwinian approach, this time to the
study of mind. There the pairing is made between the
unconscious mind and consciousness. These mental
constructs support introduction of a consciousness
thesis and then a notion of a dynamic self-referential
grammar that generates a growing repertoire of con-
sciousness manifestations. To support this approach,
an analytic set theoretic method is used to show how
a neuronal state generates a specific concept (thema)
and an example, namely the concept of the natural
numbers, is constructed. The approach suggests
experimental arrangements for study of the connec-
tion of behavior and human information processing.
Three classes of such experiments are described in
the following section.
Experiments
I offer a list of possible experiments associated to the
ideas and issues raised.
Referring to the consciousness thesis described
earlier: A winning thema could be determined
experimentally by methods that characterize the
neural correlate of a particular conscious per-
ception. Refer to the literature on experimental
determination of neural correlates (Krekelberg,
Dannenberg, Hoffmann, Bremmer, & Ross,
2003; Kato et al., 2009). Connections between
themata and neural correlates were discussed
earlier in this article. Behavior and neural corre-
lates are well documented by experiment (Kim
& Shadlin, 1999; Jeannerod, 1994).
316  •  miranker
This content downloaded from 70.95.156.2 on Sat, 02 Jul 2016 04:24:56 UTC
All use subject to http://about.jstor.org/terms
Referring to the grammatical thesis for con-
sciousness described earlier: The existence and
functioning of this intrinsic grammar could be
studied along the lines of the experimental work
on Chomsky’s universal grammar for language
development. Refer to the large body of such
work, an example of which is in Crain and
Thornton (2000); also see Coswell (2007).
This article mentions a number of open ques-
tions concerning evolution and consciousness,
each of which points to further study of de-
velopment, and in some cases to experimental
investigation. See Ben Jacob et al. (2004) for an
example of primitive behavior and cognition
(memes).
Notes
Address correspondence about this article to Willard L.
Miranker, Department of Computer Science, Yale University,
New Haven, CT 06520-8285 (e-mail: Willard.Miranker@
yale.edu).
1. Phenotypes are typically described as a behavioral
state (Blackmore, 1999), and neural states are described as
endotypes, or intermediate steps between genes and their
phenotypes (Caccioppo & Decety, 2009; Canli, Hooker, Ger-
mine, Knight, & D’Esposito, 2006; Caspi & Moffitt, 2006).
Thanks to an anonymous referee for these observations. For
convenience, I shall not use the term endophenotype but the
simpler phenotype.
2. Both genetic replication and Hebbian synaptic dynam-
ics are subject to disturbances. Mutation is the term given to
the former and noise to the latter. Of course, genes may be
exchanged in reproduction (crossover). Genes are known to
change their location within a DNA sequence.
3. One recognizes this point of view as an aspect of con-
temporary embodied cognition psychological research. This
view dates back at least to Kant (1781): “All our knowledge
begins with the senses, proceeds then to understanding and
ends with reason. . . . It is beyond a doubt that all our knowl-
edge begins with experience.”
4. Recall that a primitive is an undefined construct, de-
scribed by the properties that are required of it.
5. A study of self-reference is framed in terms of non–
well-founded sets. See Aczel (1988) and Miranker and Zuck-
erman (2009).
References
Aczel, P. (1988). Non–well-founded sets. CSLI Lecture Notes,
14.
Belkin, N. (2003). Creutzfeldt–Jakob disease: Identifying pri-
ons and carriers. Journal of the Association of Periopera-
tive Registered Nurses, 78(2), 204–210.
Ben Jacob, E., Becker, I., Shapira, Y., & Levine, H. (2004).
Bacterial linguistic communication and social intelligence.
Trends in Microbiology, 12, 366–372.
Blackmore, S. (1999). The meme machine. Oxford, England:
Oxford University Press.
Caccioppo, J., & Decety, J. (2009). Loneliness affects how the
brain operates. Science Daily.
Canli, T., Hooker, C., Germine, L., Knight, R., & D’Esposito,
M. (2006). Amygdala response to facial expressions re-
flects emotional learning. Journal of Neuroscience, 36(35),
8915–8922.
Caspi, A., & Moffitt, T. (2006). Gene–environment interac-
tions in psychiatry: Joining forces with neuroscience.
Nature Reviews. Neuroscience, 7, 583–590.
Chalmers, D. (1996). Conscious mind: In search of a funda-
mental theory. Oxford, England: Oxford University Press.
Chomsky, N. (1953). Systems of syntactic analysis. Journal of
Symbolic Logic, 18(3), 242–256.
Coswell, D. (2007). Chomsky for beginners. New York, NY:
Writers and Readers.
Crain, S., & Thornton, R. (2000). Investigations in universal
grammar: A guide to experiments on the acquisition of
syntax and semantics. Journal of Linguistics and Philoso-
phy, 16(1023/A), 523–532.
Darwin, C. (1859). On the origin of species. London, England:
John Murray.
Dawkins, R. (1999). The selfish gene. Oxford, England: Ox-
ford University Press.
Edelman, G. (1978). The mindful brain. Cambridge, MA:
MIT Press.
Fogel, D. (2000). Evolutionary computation: Towards a new
philosophy of machine intelligence. New York, NY: IEEE
Press.
Haykin, S. (2008). Neural networks: A comprehensive founda-
tion (3rd ed.). Upper Saddle River, NJ: Prentice Hall.
Hebb, D. (1946). On the nature of fear. Physiology Review, 53,
259–276.
Hebb, D. (1949). Organization of behavior: A neurophysiologi-
cal theory. New York, NY: Wiley.
Jeannerod, M. (1994). The representing brain: Neural corre-
lates of motor intention and imagery. Behavior and Brain
Science, 17(2), 187–224.
Kairiss, E., & Miranker, W. (1998). Cortical memory dynam-
ics. Biological Cybernetics, 74(4), 277–292.
Kant, I. (1781). Critique of pure reason. New York, NY: Co-
lumbia Press.
Kato, J., Wede, H., Kabashima, I., Kadota, H., Takano, K.,
& Kansaku, K. (2009). Neural correlates of attitude
change following positive and negative advertisements.
This content downloaded from 70.95.156.2 on Sat, 02 Jul 2016 04:24:56 UTC
All use subject to http://about.jstor.org/terms
Frontiers in Behavioral Neuroscience. doi:10.3389/
neuro.08.006.2009
Kim, J., & Shadlin, M. (1999). Neural correlates of a decision
in the dorsolateral prefrontal cortex of the macaque. Na-
ture Neuroscience, 324(5928), 759.
Koch, C. (2004). The quest for consciousness: A neurobiologi-
cal approach. Denver, CO: Roberts.
Koch, C., & Crick, F. (2001). On the zombie within. Nature,
411, 893.
Krekelberg, B., Dannenberg, S., Hoffmann, K.-P., Brem-
mer, F., & Ross, J. (2003). Neural correlates of implied
motion. Nature, 424, 674–677.
Lumsden, C., & Wilson, E. (1981). Genes, mind and culture:
The coevolutionary process. Cambridge, MA: Harvard
University Press.
Margulis, L. (1991). Symbiosis as a source of evolutionary in-
novation: Speciation and morphogenesis. Cambridge, MA:
MIT Press.
Miranker, W. (2005). Consciousness: A Darwinian process.
Yale University DCS TR 1344.
Miranker, W. (2010). Modeling consciousness as an adaptive
phenomenon. Quarterly of Applied Mathematics, 68.
Miranker, W., & Zuckerman, G. (2009). Mathematical foun-
dations of consciousness. Journal of Applied Logic, 7(4),
421–440.
Miranker, W., & Zuckerman, G. (2010). Dynamics of mental
activity. Journal of Applied Logic, 8(1), 114–140.
Moschkovakis, Y. (2005). Notes on set theory. Berlin, Ger-
many: Springer.
Sacks, O. (1995). Anthropologist on Mars: Seven paradoxical
tales. New York, NY: Knopf.
Sacks, O. (2007). Musicophilia. New York, NY: Random
House.
Sapir, E. (1921). Language: An introduction to the study of
speech. New York, NY: Harcourt, Brace.
Selfridge, O. G. (1959). Pandemonium: A paradigm for learn-
ing. In The mechanisation of thought processes (pp. 511–
599). London, England: Her Majesty’s Stationery Office.
Semon, R. (1921). The mneme. London, England: Allen &
Unwin.
Sheehan, E. (2006). The mocking memes: A basis for auto-
mated intelligence. Bloomington, IN: Authorhouse.
Sheets-Johnstone, M. (1998). Consciousness: A natural his-
tory. Journal of Consciousness Studies, 3, 260–294.
Su, X., & Wellems, T. (1996). Toward a high resolution Plas-
modium falciparum linkage: Polymorphic markers from
hundreds of simple sequence repeats. Genomics, 33, 430.
Wilson, E. (1999). Consilience: The unit of knowledge. New
York, NY: Knopf.
memes and their themata  •  317