The construction of ignorance and the evolution of knowledge
(Working Paper)

P. Bouissac
paul.bouissac@utoronto.ca
University of Toronto

	During the last few decades, scientific research conceived as the
empirical pursuit of truth and objectivity has been put on trial.  For
many, scientific knowledge has lost its prestige, integrity and quasi
transcendence, and has come to be variously considered as a set of rival
incommensurable theories similar to ideologies and religions, unstable
configurations of power and authority prone to irrational changes, or even
fabulations, fictions generated by self-delusion, if not deliberate
deception.  Landmark works in this critical process include Kuhn (1970),
Popper (1972), Feyerabend (1975) and Lakatos (1978).  Their criticisms have
triggered many debates (e.g. Suppe 1977, Serres 1989) and have permeated
our "épistémé" in more or less radical forms.1  Scholars from all fields
have picked up the momentum and transferred these arguments from the
natural to the social sciences, and, more generally, have questioned any
intellectual endeavour inspired by the modern scientific ethos or modelled
on the empirical method.  The notion of knowledge as revealed or discovered
truth has shifted toward a view which holds that knowledge is constructed,
if not fabricated.  Trivial versions of this new scepticism have filtered
through the general population via the media, and, according to some,
explain the marked decline in government funding for scientific research
which characterized the 80s.  In an article-manifesto, prominently
published in Nature, two physicists, T. Theocharis and M. Psimopoulos
(1987), called the attention of the scientific community to this state of
affair, noting that the crisis of confidence had reached even the core of
the scientific establishment itself and that this fact could be ominous.
They list the dangerous consequences which this critical movement might
entail for society.  Even if some aspects of their counterattack, which
blends theoretical and pragmatic arguments, may sound naive at times, their
warning should not be taken lightly, since,indeed, "scepticism and nihilism
were not threats only in the remote past; nor do they now constitute a risk
of only academic interest" (598).  Confronting  the issues raised by this
epistemological debate is not only an intellectual challenge, but it also
has political and ethical dimensions.  Efforts toward a critical
understanding of the scientific development of knowledge, including its
sociological and rhetorical strategies, do not necessarily lead to a
radical dismissal of its value and virtue.  It may, however, help reassess
its significance in the broader theoretical context formed by the
interfaces which are presently developing between the humanities and the
social sciences on the one side, and the formal and the natural sciences on
the other.

1.	Objects of perplexity
	Knowledge has always an object.  Prior to becoming an object of knowledge,
was this object an unknown entity or agency?  Is it instead a fallacy, a
total or partial creation of knowledge itself?  Are descriptions and
explanations only means toward the fabrication of world views which
legitimize political economies?  Are experience and experimentations mere
delusions, or do scientific methods of investigation yield the knowledge of
universal laws which account for the qualities, quantities and behaviours
of particular objects of perplexity which entice humans' minds into a quest
for their knowledge?  Or, are scientific institutions ways of building
fictive objects on a grand scale, inventing problems and thriving on tax
payers' money?2  Or is science the patient and eventually successful
confrontation of human intelligence with elusive objects, functions and
laws whose discovery gives earthlings the power to act upon their
environment and exactly predict the outcome of their actions?
	Many of these questions have, at one time or another, inspired
philosophical debates and political controversies over approximately two
millenia of epistemological anxiety.  They become all the more pressing in
times of scarcity now that the scientific institutions specializing in the
production of knowledge have developed close associations with military and
economic power structures; contemporary society cannot afford to make hasty
decisions because the stakes are so high that acting lightly in such matter
amounts to playing Russian roulette.  This is why, beyond the pure pleasure
of theorizing, a reflexive enterprise focusing on these problems is in
order.
	Obviously, the first relevant question to ask bears upon the status of the
object of knowledge.  This issue has triggered intense discussions
particularly in relation to experimental research (e.g., is the object an
artifact?) and to the social sciences (e.g., is the object a function of
the position of authority of the investigator?).  Some problems in the
latter domain will be now examined as a starting point for the argument I
want to develop in this article.
	One may ask for instance: Which method is most appropriate for reaching an
accurate description and a satisfactory explanation of a socio-cultural
object?  This question suggests that, on the one hand, institutions and
their productions stand as objects and that, on the other hand, methods are
commodities among which a discriminating researcher can select the one
which is best suited to the task at hand, given the assumed nature of the
object.  But such naive positivism tends to ignore the extent to which
methods and objects are intimately associated, almost indistinguishable
from each other like mutually definable correlates.  Indeed, any method
presupposes a model of its object, conceived in terms of the sort of
information that the method is conceptually equipped to detect, if not to
construct.  Methodological strategies necessarily imply the elaboration of
their object of inquiry within a more general conceptual scheme, and
endeavour to determine the range of variables set forth by the model that
is thus elaborated.
	For example, analysing public performances in general as acts of
communication presupposes a pre-determined range of variables and their
relations combined in a model.  The model is the hypothesis which
constrains the method as much as the method sets limits on the set of
admissible models.  The performance-as-communication model already contains
the heuristic categories of commutable elementary units, syntactic
structures and discursive strategies that the method will specify in
accordance with the principles of functionalist and structural linguistics
and its theoretical extrapolations.  In other words, the method makes the
implied model of its object explicit and determinate, a process which may
create the illusion that an autonomous domain of inquiry has been fully
explored and revealed as a result of the judicious application of
analytical procedures of a particular sort to some `raw' material.
	Is this to claim that all methods are equally valid or equally vain, and
that the quest for knowledge proceeds through a forest of epistemological
fantasies?  Are researchers coming from various theoretical horizons doomed
to confront one another in a conflict of interpretations?  Is the cognitive
interplay between models and methods necessarily condemned to be circular,
thus locking researchers, individually or as groups, in epistemological
solipsism? This is, of course, a possibility, and, as pointed out above,
contemporary polemical debates, focused upon epistemological relativism,
have called attention to the extent to which theories, scientific or
otherwise, are parts of larger social, economical, cultural and ideological
contexts which  change over time under the pressure of sets of conditions
which are so complex that they may appear at times as random and irrational
processes.
	However, the constant shifting of models or conceptual schemes does not
suffice to justify the radical epistemological relativism of those who hold
a purely `historicist' view of knowledge (D'Amico 1989).  It can be
expected indeed that empirical information always exceeds the formal object
of an inquiry; the method can perform its sifting and organizing functions
only to a certain extent.  The case can be made that there is a threshold
beyond which the researcher can no longer ignore the accumulation of
residue or noise.  In spite of their inner beauty and logical consistency,
models are eventually overcome by their insoluble surplus.  Intellectual
probity or a mere concern for pragmatic efficiency then motivates the
restructuring of the initial model or the constituting of the residue
itself as an object of inquiry with its own correlative model and method.
When the descriptive and explanatory power of a model is strained by the
number of exceptions to the rules it generates or by the shrinkage of the
domain of experience to which it applies, other attempts at modelling are
undertaken.  The challenge, however, remains to conceptualize the
relationship among alternative or successive models, especially if it
appears that the production of noise results from the very nature of
models.  The belief that socio-cultural phenomena can be exhaustively
represented in terms of a unified theory may help one sustain the effort
required to conduct systematic inquiries; but that remains a utopian goal.
Heuristic blindness can be productive only to a certain degree.  The crisis
of representation has brought to the awareness of many researchers the
inherent artificiality not only of representational devices, but also of
the objects which are constructed through representation along the
investigative process, and which can acquire an inhuman dynamic of their
own.  This awareness can indeed take the form of a deep moral crisis when
it turns out that human agents are thus objectified as `others' and
ultimately construed as mere tokens in some academic power game based on
social predation.  Ethnographic information can be indeed gathered as a
mere commodity without consideration for the political conditions which
make such an endeavor possible or for the consequences it may have locally.
The reflexive anthropology of the 1980s (e.g. Clifford and Marcus 1986,
Marcus and Fischer 1986) has forcefully voiced both ethical and
epistemological problems that cannot be glossed over and are made all the
more pressing at a time when anthropological pursuits, driven by the
pressure of evolving geo-political circumstances, focus on the
anthropologists' own society, and encounters most revealing challenges
(Jackson 1987).
	But where does epistemological despondency lead if not to a dark age of
anomy, both intellectual and social?  In response to the predicaments of
knowledge, one may want to devise less arrogant approaches than the one
with wich positivism is usually associated.

2.	The spectre of the object
	It is possible that the intractable problems resulting from a single,
exclusive focus partly dissolve when frames and roles are systematically
shifted.  This alternative is at least worth exploring.  The resistance to
the constructed objects which manifests itself in the form of parasitic and
haunting noise and the subsequent drifting of theoretical elaborations
suggest some sort of refraction which calls for a spectroscopic approach.
Thus, hypothetical objects are not decomposed in terms of the analytical
units or functions of a single model, but instead the targeted domain of
experience is reconstructed in terms of differential, serial, and
interfacial theoretical objects and subjects.
	Interfaces may be relative, arbitrary, artificial and discrete to a
degree, but they are nevertheless bounded within a continuum limited only
by the horizons of experience.  If, furthermore, these horizons are
conceived as defined by the experience of a community of `researchers' and
`researched' rather than by an individual's or school's point of view, a
more comprehensive object of knowledge can be constructed that is at the
same time complex, ordered and real.  But before reaching this intellectual
and moral grasp of an object that is both diverse and consistent, it is
necessary to construct a multiplicity of interfaces.  This approach might,
of course,raise the spectre of a fractal infinite regression, except that
it does not intend to be exhaustive and can easily admit that interfaces
are, to a degree,  historically conditioned as the trans-object itself
necessarily is.
	A way of conceptualizing the complexity of a socio-cultural phenomenon as
a domain of experience is indeed to model it as an array of interfaces
without any suggestion of a principled hierarchical system. An interface is
usually defined as a surface forming the boundary between adjacent spaces
of whatever nature and consistency these spaces may be.  Technically this
surface equally belongs to the two contiguous domains but has also some of
the properties characterizing each.  The concept of interface can be
further abstracted and recast in semiotic terms as the plane of contact
between two systems modifying each other through some form of symbolic
interaction.  The contact between a liquid and a solid, which is one of the
most striking examples of an interface, is only a particular case.  There
are also interfaces between magnetic fields as well as between deictic
systems of signs such as those defining the boundary between two
ethological territories.  In fact all communication processes imply an
interface between two or more information sources or fields. Interfaces may
indeed overlap and probably necessarily do so.  Social groups, marked
spaces, fences, social distances, face-to-face interactions, spectacles
form interfaces whose correlated and synchronized modifications can be
described as long as these interfaces are temporarily and heuristically
construed as closed systems.  All social realities can be seen as
structured, or rather textured, by complex systems of interfaces on various
levels of inclusiveness.
	When particular interfaces are constituted as objects of inquiry,
configurations of factors are put in place and hierarchical, or at least
structural, orders define the objects of inquiry and determine their
method.  Both descriptive procedures and explanatory interpretations can
then be developed in a manner congruent with the interfaces that have been
thus constructed.
	However, it must not be forgotten that artificiality, or rather
`artifactuality', belongs indeed to the nature of interfaces.  It would be
misleading to conceive of interfaces as `slices' cut through the `real' of
the institutions and their productions, and closely examined with a
magnifying device as if they were biological tissue. The construction of an
interface involves, on the one hand, the recognition of a zone of contact
between two distinct domains and its methodological isolation and, on the
other hand, the formulation of a hypothesis or model which determines which
sort of information will be collected and organized as being constitutive
of these conceptualized zones.  This approach basically relies on a
theoretical systematization of the natural cleavages of socio-cultural
experience with respect to societies, groups, generations, sub-cultures,
corporations, specialties, and so on.  But these cleavages are so numerous,
complex, multivalent, fractal, fluctuating and chaotic that any
methodological determination of an interface can only be a grossly
simplified model.  However, building series of such interfaces may bring
some measure of systematicity and coherence to the notion of `thick'
description which Geertz (1973) borrowed and adapted from Ryle (1971) and
which seems to be the only possible response to the engrossing,
overwhelming object that any socio-cultural phenomenon form as a domain of
experience.
	Thus, one may be led to conclude that it is not because an object of
inquiry is constructed that the research process does not yield any amount
of objective knowledge, as long as the construction process itself is taken
into account and forms an explicit part of the knowledge thus acquired.
However, this perspective, which may help clarify the status of objects of
knowledge as processed domains of experience remains largely tautological
and strictly pragmatic.  A further questioning of the strategy atplay in
this process may contribute to a better theoretical understanding of the
epistemological problems which are being addressed in the current crisis.

3.	An architecture of holes, gaps and lacunae
	Some fifteen years ago, Duncan and Weston-Smith (1977) edited a dense
volume of collected articles under the tantalizing title The Encyclopaedia
of Ignorance.  Prominent specialists - from Great Britain for the most part
- in a wide array of scientific domains had been invited to write about the
yet unanswered - or unanswerable - questions in their respective fields of
inquiry.  The resulting texts provide a most enlightening vista on the
process through which knowledge is constructed in the sciences.  Far from
radiating unlimited confidence in their ability to conquer the unknown as
if it were an object to be mapped, measured and packaged, the contributors
to this volume carefully delineate their own area of ignorance and
emphasize the crucial part that this process plays in their pursuit.  Each
one relies on a particular rhetorics to express this essential absence of
knowledge.  For example, some evoke a daunting horizon of increasing
complexity, some wonder whether the human brain can cope at all with this
complexity beyond the requirements of extended survival strategies, some
even question the very possibility of knowledge.
For instance, Lehman, a computer scientist,  declares in conclusion of his
contribution: "Total knowledge, the final state, can never be reached.
Ignorance must always be present" (354).  This should be understood not as
a romantic declaration but as a technical, almost clinical constatation.
It raises an interesting question regarding cognitive and discursive
strategies through which scientific knowledge is constructed.  Usually, the
cumulative process of science is metaphorically apprehended as a
`positive', ever increasing sum.  However, it also could be conceived as
the `negative' - in the photographic sense of the term - of an activity
geared toward the relentless construction of ignorance, an architecture of
holes, gaps and lacunae, so to speak.  This might sound paradoxical, but
The Encyclopaedia of Ignorance reveals a consistent strategy which consists
in conceptualizing what is not known, not in a vague, open-ended manner,
but with application, precision, almost industriousness.  It appears that
the excercise did not seem strange to these scientists for the good reason
that setting up what could be called "stages of uncertainty" is the very
stuff of their professional existence.  If, indeed, knowledge is
information, its condition of existence and its real measure is
uncertainty, and therefore depends on the construction of ignorance.  There
is no possibility of information if there are no alternatives with various
degrees of probability.  The more unexpected, the greater the discovery
because the uncertainty is resolved by a choice which was not even
seriously taken into consideration when the most probable alternatives were
laid out in the experimental design.  Such discoveries suddenly cause a
reframing of ignorance because alternatives previously considered to be
equally probable in view of a given knowledge become improbable and, thus,
the presupposed knowledge which had led to this state of affair must be
retroactively discarded and replaced by new questions which redefine the
horizon of ignorance.3
	Naturally, "science" or "scientific knowledge" as an entity is a
rhetorical device.  A grammatical subject which obfuscates the politics of
institutions, somewhat like "life" covers the swarming and teeming of
myriads of competing organisms.  Academic disciplines themselves are
dramatis personae on the political stages which appear whenever a critical
mass of researchers agree on the definition of a domain of ignorance and on
`legitimate' methods to perpetuate this ignorance by a constant renewal of
the stage of uncertainty.  It should be clear at this point that
`ignorance' is not taken here in the absolute sense of self-ignored
ignorance, i.e. an absence of knowledge caused by theabsence of a stage of
uncertainty in the representational system of an organism.  The ignorance
to which it is referred here is more akin to the state of knowledge of a
sleuth whose reasoning is portrayed at the beginning of a crime fiction
than to the situation of someone who would have been always immersed in
total darkness.  Scientific disciplines need decidable uncertainties but
need them for ever, so that the narrative will continue.  State of the art
articles as a genre always are structured by what could be called a dynamic
of lack.  The Khunian paradigms need solvable but indefinitely expanding
puzzles.  In as much as these puzzles are expressed in the form of
discourses, they are rhetorical artifacts.  But, by contrast with the
Humanities and some sectors of the social sciences which thrive by staging
undecidable uncertainties, the empirical sciences construct decidable ones
by specifying in each case what will count as information, and which new
ignorance this information will help formulate.  From a general point of
view, it seems impossible to conceive of a discipline which would consist
of the mere transmission of a completed knowledge.  A discipline which
would fail to generate information, i.e. which would have exhausted its
capacity to construct ignorance, would quickly disappear.  The necessary
dynamic must come from the construction of a relatively determined lack of
knowledge.  Therefore disciplines can be viewed as generators of
uncertainty and can be compared with each other from this point of view.
As it has been hinted above there are many forms of ignorance depending on
the generator's model.  Each discipline seems to be based upon a different
blueprint, or a program (in the computer sense of the term) which specifies
how to construct its sui generis uncertainties.  But such programs are
necessarily in competition with each other because they need hardware (i.e.
human and economic resources) in order to be implemented.
	An illustrative case is semiotics, a discipline-to-be which did not
succeed in achieving disciplinary status, although it developed a dynamic
toward this goal.  It is symptomatic that the two `founding' programs, i.e.
those of Peirce and Saussure, essentially consist of discursive fragments
about some unknown to be discovered; both evoked a relatively well
specified horizon of ignorance but fell short of determining with optimal
quality the generative model which could have yielded a self-replicating
puzzle.  As a result, the stage of uncertainty which defines semiotics
remained populated by undecidables which can generate indefinite discursive
productions, but finds themselves in competition with more powerful
programs of uncertainties such as those of philosophy and literary
interpretation.  This state of affairs is appropriately expressed through
the convoluted rhetoric of undecidability in the concluding paragraph of an
essay by a representative semiotician: "A final remark, which is for
obvious reasons inconclusive: From the preceding propositions one might get
the impression that this program of enquiry will result, all in all, in a
referential tautology.  Well, it could also be said - precisely!  Giving
due regard to what M. Eigen (1975) has stated at the outset of his
influential article on "Evolutionary Games", let us take in, henceforwards
[interdiscursively] the basic rule of reconstruction as he puts it: "The
origin of life [intertextuality] is tautologous with the origin of
biological [textological] information."  Therefore, ours the task, may be
eternal." (Ruprecht 1991:73)
	It is, indeed, undoubtedly so, but by design rather than because of an
epistemological curse, as post-modern romanticism would have us believe.
Naturally, "design" is taken here in the sense of "structural feature", and
not psychological intentionality.
	The same process of constructing a form of ignorance is observable across
the whole array of modern disciplines which are thus provided with the
dynamic needed for evolving and persisting over time.  But each discipline
is defined by its particular rhetorical strategies and modalities of
implementing the perpetual reconfiguration of its stage.
	At this point, one may be inclined to conclude, as some do, that knowledge
production is a mere rhetorical trick, that special terminologies are ways
of obfuscating tautological statements, and that the notion of scientific
progress is nothing but an ever changing narrative used to legitimize an
otherwise frivolous game of power and confrontations.  However, an argument
to the contrary can be extrapolated from evolution theory - not as a
dogmatic derivation but as a way to account for the fact that, indeed, some
forms of constructed ignorance and the information it yields often make a
difference for the organisms engaged by this process.
	Let us recall that evolutionism is, briefly, the theory which holds that
selection operates on random variations. It is a non-teleological process.
Adaptation is not conceived as a goal-oriented change but as a
serendipitous advantage yielded by a particular transmissible modification
of an individual phenotype.  For an advantageous variation which
contributes to evolution, there are countless detrimental ones leading to
early elimination of those genotypes from the genetic pool. Moreover, what
is advantageous in a particular environment, can become a liability in a
different one or if the initial environment changes.  At the cost of a
metaphoric translation, strategies which elicit information can be
conceived as phenotypes implementing genotypic variations whose outcome
may, at times, confer a decisive advantage through a close fit with
environmental constraints.  There are actually no good reasons for
excluding rhetorics from the domain of natural phenomena, nor for
explaining it in merely functionalist terms.
	But, having set the theoretical perspective which will be further
discussed in the next section, let us return for a while to the strategic
construction of ignorance and let us examine what `constructing ignorance'
practically means.  It consists of building a model, generally of an
interface, such that some relations constitutive of the model remain
indeterminate.  For instance, in Duncan and Weston-Smith (1977), Barlow, a
specialist of physiological optics, addresses the topic of the languages of
the brain: "The language for communication between the various higher parts
of the brain must use symbols with an agreed meaning, for otherwise one
part would not understand what another part said.  We do not know what
these symbols are, nor how their meanings are agreed to.  However, the
material discussed in this essay does enable us to hazard a few guesses
about the general nature of the language, and here are three of them. (1)
Whatever a symbol may correspond to in the external world, internally it is
likely to be represented by nerve impulses in a nerve cell or in a family
of nerve cells, for we know of no other tokens that can be rapidly
exchanged between the parts of the brain.  (2) There is nothing to indicate
that these internal symbols occur in a single serial string [...]"
(271-271),  This suffices to make clear that the model construes "part of
the brain" as agencies which interact through communication, i.e. transfer
of information; while there is evidence that such interactions occur, the
mode of communication is not known.  This uncertainty is, however, reduced
to a few alternatives which are introduced as the stage of uncertainty
within the model itself.  The narration is familiar: the sleuth has
evidence that two conspirators communicate through a secret code.  Since
they live in different cities, the messages have to be carried.  By whom
and in which form are these messages transmitted? and so on.  This was the
state of the art fifteen years or so ago.  Neurotransmitters are now better
known, but this new knowledge creates its own horizon of uncertainty.

4.	Epistemology: from creationism to evolutionism
	It seems that, at this point of the reflexion, at least two assertions can
be stated which would be supported by a reasonable consensus.  First, the
process of scientific observation necessarily  implies some form of
theoretical presuppositions, hence the logical pre-eminence of theory over
scientific observation and experimentation.  This argument has been put
forward and convincingly argued by Popper (e.g. 1972:340-361) and the
counter argument that theories can be made obsolete by disproving data does
not establish the reverse.  It simply shows that a theory can be replaced
by another.  This leads to the second assertion, namely that theories are
entities in competition with each other.  Some are short lived or remain
confined to a small circle of minds; others spread among much larger groups
of thinkers, researchers, and teachers; some gain the status of worldviews,
or commonsense, by taking hold of whole populations.  The paradigm shifts
described, but not explained, by Kuhn, dramatizes this eliminative process
by evoking some "irrational" forces.
	If it is true that, as has been suggested by the argument developed in the
second and third sections of this essay, theories are differenciated ways
of constructing uncertainty, and correlatively, of producing information,
it could be concluded that they share an important, if not critical,
feature with living organisms.  This line of argument draws from current
assumptions made in artificial life (AL) research, holding that the
"logical form" of an organism can be separated from its material basis of
construction, and that "aliveness" will be found to be a property of the
former, not of the latter" (Langton, 1991:11).  The way, in which theories
come and go, prosper and disappear, indicates that, although they are
generally considered to be produced by human brains, they are endowed with
a relative degree of autonomy with respect to the populations of organisms
among which they spread with various degrees of success.  Traditional
philosophies and religions have developped an ontology of ideas,
narratives, instructions and rituals which take such autonomy for granted
(e.g. Plato's ideas; divine words embodied in sacred texts; rituals
identified as divine beings in Hinduism).  Moreover, secular poets,
philosophers and scientists have often referred to the sui generis dynamism
of ideas or theories as quasi agencies independent from their own psyches.
Metaphorical expressions in many languages emphasize this puzzling
impression: one comes across an idea, one may yield to the power of a
worldview or a system, one may be inspired, possessed, turned on by a
theory, and so on.  Considering the way in which myths seem to unfold and
transform their structures as they drift across continents from population
to population, Levi-Strauss once noted that "The Ojibwa Indians consider
myths as `conscious beings', with powers of thought and action" (1969:12,
note) and claimed that his goal was "to show not how men think in myths,
but how myths operate in men's minds without them being aware of the fact.
(ibid:12)  Someone can also identify with a particular theory and
passionately fight for it - many humans have died for a worldview and still
do - and, some times, can as easily switch to another one without
noticeable detrimental effect on his/her organism and mental life.  Ideas,
faith, commonsense, belief can be shown to share features with theories,
even to be theories under other names.  Theories have no particular regard
for the welfare of the populations among which they may thrive .  They may
even conceivably go extinct because "their" population disappears,
precisely because of some ill effects of the normative behaviour they
consistently promote in the population.  On the other hand, some theories
may have some long term beneficial side effects on "their" populations.
But it does not seem that as an information structure a theory as such
could have either a benevolent or malevolent intent.  It simply spread as
fast and as much as possible, eliminating competing theories in the
process, sometimes softly by causing them to be forgotten, sometimes
violently by physically suppressing the populations which hold them.
	Naturally an extended metaphor cannot be used as an argument.  Common
sense associates theories so closely to the mental lives of their
generators that such evolutionary analogies probably will be considered as
nothing but playful speculations.  However, it could also be pointed out
that this perceived indissociable association, this identity of an
information-producing structure and the brain which implement it, might be
just a theory among others.  But, then, it might be opposed to this line of
thought that, according to a well known principle, one is not entitled to
demonstrate a theory by this theory itself, except of course if this new
theory had such a powerful information producing structure that it would
"entropize", so to speak, Gödel's argument.
	The notion of evolutionary epistemology goes back to the end of the XIXth
century.  But its development remains anthropocentric.  The survival rate
of theories depends on their usefulness to humans; in fact they are closely
related to the fitness of particular individuals.  They are part of the
phenotype.  Popper, however, who discussed the issue on several occasions
(e.g. 1972:65-70) perceived the inherent problems attached to this view:
"[...] questions of truth or validity, not excluding the logical
justification of the preference for one theory over another (the only kind
of `justification' which I believe possible) must be sharply distinguished
from all genetic, historical, and psychological questions. [...] logical
investigations of questions of validity and approximation to truth can be
of the greatest importance for genetic and historical and even for
psychological investigations.  They are in any case logically prior to the
latter type of question, even though investigations in the history of
knowledge can pose many important problems to the logician of scientific
discovery". (1972:67-68) (emphasis in the text).
	This is a way of saying - albeit from another theoretical viewpoint than
the one proposed here - that organisms and theories are entities of
different order.  But the order which is constitutive of Popper's theory
implies a hierarchy whereas "order" will be taken here in its taxonomic
sense.
	Although the evolutionary selection metaphor applied to theories appears
often in the epistemological literature of the last two decades, it is not
until the bold, albeit tentative introduction of the concept of "meme" by
Dawkins (1976:206) that a new way of thinking evolutionary epistemology
could be proposed.  Let us recall the initial, somewhat awkward formulation
of this hypothesis: "Just as genes propagate themselves in the gene pool by
leaping from body to body via sperms or eggs, so memes propagate themselves
in the meme pool by leaping from brain to brain via a process which, in the
broad sense can be called imitation.  If a scientist hears, or reads about
a good idea, he passes it on to his colleagues and students."  The term
"meme" was somewhat playfully coined after the French "même" for "same" and
with phonetic allusions both to "mimetic" and to "gene".  In the paragraph
following this introduction of the new concept, Dawkins quoted the summary
that one of his colleagues, N.K. Humphrey, produces in reaction to the
reading of an earlier draft of this chapter: "memes  should be regarded as
living structures, not just metaphorically but technically.  When you plant
a fertile meme in my mind you literally parasitize my brain, turning it
into a vehicle for the meme's propagation in just the way a virus may
parasite the genetic mechanism of a host cell.  And this isn't just a way
of talking - the meme for, say, "belief in life after death" is actually
realized physically, millions of times over, as a structure of the nervous
systems of individual men the world over." (207)  In his subsequent
publications Dawkins returns to the concept of "meme" (1982: 97-117,
1986:158) and indicates attempts made by others to develop his hypothesis
(e.g. Delius, 1990) which has also been taken up since by Dennett (1991).
	The idea of entities endowed with both genotypic and phenotypic structures
independent from the biological ones but evolving in a parasitic relation
to them, remains an intriguing and fascinating metaphor which shatters our
world view as if it were an aftershock of Darwinism but is too "odd" to be
taken seriously.  Dawkins himself has been cautious in his subsequent and
brief treatments of the "meme" idea, as if he were daunted by the full
implications of this emergent theory.  It is worth noting in passing that
"memes" are strikingly different from the "culturgens" conceived by Lumsden
and Wilson (e.g. 1981:27, 368) which remained modelled on functional
artefacts and the process of cultural inheritance.  Even Boyd and Richerson
(1985) for whom "culture is a system of inheritance [...].  Today's
cultural traditions are the result of cumulative changes made by past and
present bearers of them" (20), shy away from the full implications of
Dawkins' views.  They state: "Our definition of culture is not at all
specific about the nature of the information that affects phenotypes.  In
particular, we do not assume that culture is encoded as discrete "particles
[...].  Relatively little can be said on this topic since our knowledge of
the neurophysiology of social learning is primitive compared to our
knowledge of the molecular biology of the gene" (37).  However, when they
address the issue of "horizontal" transmission, i.e. transmission within a
generation rather than across generations, the analogy with parasitic
phenomena naturally comes to their minds: "Horizontal transmission is
analogous in some ways to the transmission of a pathogen [...]  The item of
culture being spread horizontally acts like a microbe that reproduces and
spreads rapidly because it is "infective" and has a short generation length
compared to the biological generation length of the "host.  Fads and
fashions and technical innovations are familiar examples". (8)  The
epidemiological metaphor is also used by Cavalli-Sforza and Feldman (1981)
in their quantitative study of cultural transmission.
	As long as the metaphor is constrained by the model of a biological
parasite such as a bacteria or a virus, the analogy remains tantalizing but
non-operational.  This seems to be the source of Dawkins' reluctance to
proceed further with his yet undeveloped "meme" theory.  However, another
analogy may come to the rescue.  A developing sector of computer science
focuses on information structures whose behavior within computer
environments resembles so much the behavior of viruses within a biological
milieu that the concept "virus" or its colloquial, more begnin form "bugs"
has gained wide currency, and has become the object of intensive scrutiny.
Cohen (1990, 1991) has fully explored the issue and interestingly
emphasizes that, although computer viruses have become associated with
malevolent intent in the form of criminal or mischievous artefacts, they
are, like their biological equivalent, two-edged swords: "the features that
make computer viruses a serious threat to computer integrity can also make
them a powerful mechanism for the reliable and efficient distribution of
computing resource" (199:23).  But this is not all; not only are computer
viruses subject to mutations, but also some of these mutations may happen
to be beneficial as Cohen shows with the case of the Morris' Internet
virus.  The point which seems to be most relevant here is that computer
viruses can provide an operational model for "memes" conceived as programs
endowed with their own dynamism, albeit of a non-biological nature (see
Langton 1991, mentioned earlier).  It is therefore possible to
conceptualize theories, under any other names, as information structures
(hence the centrality of uncertainty in the dynamics of their
implementation), whose relation to the environment within which they
replicate is double-edged.  In the same ways as a genetic random mutation
may by chance yield an advantage for the organism which carries it as a
part of its phenotype, an information structure variation (which can be a
copying error, a misunderstanding or a misspelling) can spread through a
population both horizontally and vertically and yield an advantage
(adaptability) for this population.  This theoretical view, which seems to
provide a first step toward a truly evolutionary theory of culture, could
explain many features of cultural and scientific evolution: why cultures
and knowledge are only relatively cumulative; why cultural traits are not
necessarily beneficial to the populations within which they thrive; why
theories constantly displace each other but sometimes make temporarily
irreversible differences in terms of the survival of the populations which
carry them.  The perspective which is thus opened might succeed in
"displacing" or "entropizing" the creationism which remains implicit in
traditional theories of knowledge, whose current crisis may represent a
conflict of evolutionary significance.


References
Abragam, A. (1985)  "Théorie ou expérience: un débat archaïque".
Le débat 34 Mars 1985 (21-31).
Bouissac, P. (1991)  "Introduction: The circus - a semiotic spectroscopy".
Semiotica 85 - 3/4 (1991), 189-199.
Brooks, D.R. and E.O. Wiley (1988)  Evolution as Entropy. Toward a Unified
Theory of Biology. Second Edition.  Chicago: University of Chicago Press.
Clifford, J. and G.E. Marcus (eds.) (1986)  Writing Culture: The Poetics
and the Politics of Ethnography.  Berkely: University of California Press.
Cohen, F. (1985)  "Computer viruses".  Computer Security: A Global
Challenge.  Amsterdam: Elsevier Science 143-159.
Cohen, F. (1987)  "Computer viruses, theory and experiments."  Computers
and Security 6:22-35.
Cohen, F. (1990)  A Short Course on Computer viruses.	:ASP Press.
Cohen, F. (1991)  "Friendly Contagion. Harnessing the Subtle Power of
Computer Viruses".  The Sciences (September/October 1991) 22-28.
D'Amico, R.C. (1989)  Historicism and Knowledge.  New York: Routledge,
Chapman and Hall.
Dawkins, R. (1976)  The Selfish Gene.  Oxford: Oxford University Press.
Dawkins, R. (1982)  The Extended Phenotype.  Oxford: Oxford University Press.
Dawkins, R. (1986)  The Blind Watchmaker.  London: Penguin Books.
Dawkins, R. (1989)  The Selfish Gene. New Edition.  Oxford: Oxford
University Press.
Delius, J.D. (1990)  "Of mind memes and brain bugs": a natural history of
culture".  In W.A. Koch (ed.) The Nature of Culture.  Bochum: Studienlag
Brockmeyer.
Dennett, D. (1990)
Duncan, R. and M. Weston-Smith (eds.) (1977)  The Encyclopaedia of
Ignorance.  Oxford: Pergamon.

Eigen, M. (1975)  "Evolutionary Games."  In I.R. Miller (ed.) Stability and
Origin of Biological Information: The first Aharon Katzir-Katchalsky
Conference [June 23-28, 1973] New York: Wiley.
Feyerabend, P. (1975)  Against Method.  London: New Left Books.
Geertz, C. (1973)  The Interpretation of Cultures.  New York: Basic Books.
Gardin, J.-Cl. et. al. (1987)  La logique du plausible. Essais
d'épistémologie pratique en sciences humaines.  Paris: Editions de la
Maison des Sciences de l'Homme (2nd edition).
Jackson, A. (ed.) (1987)  Anthropology at Home.  ASA Monographs 25.
London: Tavistock.
Kuhn, T. (1970)  The Structure of Scientific Revolutions.  Chicago:
University of Chicago Press.
Lakatos, I. (1978)  The Methodology of Scientific Research Programmes.
Philosophical Papers volume 1.  Cambridge: Cambridge University Press.
Langton, C.G. (ed.) (1991)  Artificial Life. vol. I. Redwood City (CA):
Addison-Wesley.
Lévi-Strauss, Cl. (1969)  The Raw and the Cooked.  Trans. by J. and D.
Weightman.  New York: Harper and Row.
Malamoud, C. (1989)  Cuire le monde. Rite et pensée dans L'Inde ancienne.
Paris: Editions La Découverte.
Marcus, G.E. and M.M. Fischer (1986)  Anthropology as Cultural Critique: An
Experimental Moment in the Human Sciences.  Chicago: University of Chicago
Press.
McCloskey, D.Y. (1986)  The Rhetoric of Economics.  Madison: University of
Wisconsin Press.
Nagel, E. and J.R. Newman (1958)  Gödel's Proof.  New York: New York
University Press.
Pera, M. and W.R. Shea (eds.) (1991)  The Art of Scientific Rhetoric.
Canton (MA): Science History Publications.
Popper, K.R. (1972)  Objective Knowledge. An Evolutionary Approach.  Oxford: Clarendon Press.

Ruprecht, H.-G. (1991)  "The reconstruction of Intertextuality."  In H.F.
Plett (ed.) Intertextuality.  Berlin: Walter de Gruyter (60-77).
Ryle, G. (1971)  A Collection of Critical Essays.  O.P. Wood and G. Pitcher
(eds.)  London: MacMillan.
Serres, M. (1989)
Simons, G. (1989)  Viruses, Bugs and Star Wars.  Oxford: Blackwell.
Suppe, F. (ed.) (1977)  The Structure of Scientific Theories.  Urbana:
University of Illinois Press.
Theocaris, T. and M. Psimopoulos (1987)  "Where science has gone wrong."
Nature 329 (15 October 1987), 595-598.
Thom, R. (1985)  "La méthode expérimentale: un mythe des épistemologues (et
des savants?)"  Le Débat 34, Mars 1985 (11-20).
White, J.B. (1986)  Hercules' Bow: The Rhetoric and Poetics of Law.
Madison: University of Wisconsin Press.