Naturalist
Theories of Meaning
David Papineau
To appear in E.
Lepore and B. Smith eds
1 Introduction Naturalist
theories of meaning aim to account for representation within a naturalist
framework. This programme involves two
ideas: representation and
naturalism. Both of these call for some initial
comment.
To begin with the
former, representation is as familiar as it is puzzling. The English sentence ‘
Sentences can
represent, and so can mental states. By
and large, naturalist theories of meaning take mental representation to be
basic, and linguistic representation to be derivative. Most such theories aim first to account for
the representational powers of mental states—paradigmatically beliefs—and then
to account for the representational powers of sentences in public languages by
viewing the latter as in some sense ‘expressing’ mental states.[1]
Most naturalist
theories of meaning also subscribe to some version of the ‘language of thought’
hypothesis. That is, they assume that
the vehicles of mental representation are inner items with sentence-like
structure, at least to the extent that they are constructed from recombinable
word-like components (‘concepts’) which carry their representational content
from use to use.
It is not clear
how far these commitments—to the primacy of mental representation over public linguistic
representation, and to an inner language of thought—are essential to naturalist
theories of meaning. One can imagine
versions of the theories to be discussed below that relax either or both of
these assumptions. Still, most existing
naturalist theories do work within this framework, and it will be convenient to
take it as given in what follows.
What about the
requirements of ‘naturalism’? At its
most general, naturalism says that the methods and ontology of the natural
sciences are sufficient for understanding reality. A naturalist theory of meaning would thus aim
to bring the phenomenon of representation within the scope of the natural
sciences. However, naturalism in this
general sense is a very open-ended doctrine.
There are many different branches of natural science—from physics and
paleontology to meteorology and zoology—each with its own methods and
ontologies. Without some further
specification of what counts as a ‘natural science’, it is unclear that ‘naturalism’
imposes any genuine requirements at all.
In particular, it is unclear why our everyday pre-theoretical understanding
of representation should not already qualify as naturalistic, without the help
of any further theoretical analysis.
Contemporary
naturalism normally also endorse some version of physicalism. But it is not clear that even this further
commitment imposes any substantial methodological constraints on theories of
representation. Contemporary physicalism
only requires that non-physical properties must ‘supervene’ on physical properties
(in the sense that any non-physicial differences between things must derive
from physical differences) not that they be type-identical with physical
properties (Fodor, 1974). Again, this
leaves it unclear why our everyday pre-theoretical understanding of
representation should be in need of help from further ‘naturalistic’ theorising. After all, our everyday pre-theoretical
understanding of representation already seems in perfectly good accord with the
requirement that representational facts should supervene on physical ones.
Still, even if
‘naturalism’ as such does not impose any strong reductive demands, it is not
difficult to motivate theories which aim to account for representation in terms
of such basic scientific categories as causation, spatio-temporal correlation,
functional isomorphism, or biological function.
Representational facts appear radically unlike facts found in other branches
of science. A pattern of marks on paper,
or a state in some psychological system, somehow reaches out and lays claim to
some possibly distant state of affairs.
How is the trick done? And how do
these representational relations interact with other features of the natural
world? If some theory can answer these
questions by reducing representational relations to other familiar categories,
then that would clearly constitute an achievement, whether or not such a theory
is mandated by the methodological requirements of ‘naturalism’.
From this
perspective, the proof of the naturalistic approach to meaning will be in the
eating. Naturalists will seek some a
posteriori reduction of representation to other scientifically familiar
categories, and aim thereby to show how representational relations play a role
in the scientifically described world.
If this project succeeds, then that will be its own vindication. Of course, it remains open that no such
reduction is possible. In that event,
thinkers of strongly naturalist inclinations may wish to argue that
representational relations should be eliminated from our world view, on the
grounds that nothing in reality answers to our everyday conception of
representation.[2] Others, however, will maintain that our
everyday conception of representation is acceptable in its own right, even if
no reduction to other scientific categories is possible. Fortunately, we can leave this issue open
here. Our main business is with the
prior question of whether any of the naturalistic theories so far proposed does
constitute a plausible scientific reduction of representation.
2 Inferential
Role Semantics One family of naturalist theories of meaning
take the representational content of mental states to by constituted by their inferential
role. (Harman, 1982, 1987, Block,
1986. See also Cummins, 1991, and Peacocke,
1992, for related approaches.)
Take the concept dog. This bears inferential relations to various
other concepts, including animal, mammal, and pet. Inferential role semantics takes the total
set of such inferential relations to fix the content of dog. This can be seen as involving two elements: first, the cognitive role (the connotation,
the sense) of dog is identified with this set of inferential relations; given this, the referential value (the extension,
the denotation) of dog is equated with that entity, if any, whose
real-world relations to the referents of animal, pet and so on
are isomorphic to the inferential relations dog bears to these other
concepts.
An initial problem
for any theory of this kind is to avoid conceptual holism and consequent
problems for the public communicability of concepts (Fodor and Lepore, 1992). Different subjects are unlikely ever to embed
a concept in exactly the same set of inferential relations—given my particular
views about dogs, I will no doubt infer some different things from applications
of the concept dog than you will.
If the cognitive identity of any concept depends on the totality of
inferential relations it enters into, then it would seem to follow that
different individuals will rarely share the same concept. But this seems inconsistent with the
existence of public languages, and in particular with the fact that a word
like ‘dog’ expresses the same concept in the mouths of different individuals.
The obvious
response to this problem is to say that not all inferential liaisons contribute
to the cognitive identity of concepts.
This would then allow different individuals to display idiosyncratic
inferential dispositions without this automatically rendering their concepts
incommensurable. The trouble with this
suggestion, however, is that there seems no principled way of distinguishing
those ‘analytic’ inferential liaisons that contribute to the identity of
concepts from the ‘synthetic’ ones that do not (Quine, 1951). Moreover, even if there were some way of
making this distinction, the original problem is likely to remain, for there is
no obvious reason why individuals should coincide even in those analytic
inferential liaisons that do fix the cognitive identity of concepts.
Another major
problem facing inferential role theories is the apparent circularity of the way
they explain reference. The idea is that
the referent of dog is that entity which is appropriately related to the
referents of animal, pet and so on. But what determines the referents of the
latter concepts? If their referents are
explained in the same way, as depending on the inferential relations that these
concepts bear to yet other concepts, then there would seem nothing to tie down the
overall structure of inferentially related concepts to the real world. At best that structure could be seen as
representing any set of entities that bear relations that are isomorphic
to the inferential relations between the concepts. But then it seems that dog, animal,
pet and so on will come out as representing many different
things—structures of atoms, stars, or whatever—as well as the kinds they
actually represent. For surely there are
many structures of atoms, stars, and other things that are related in ways that
are isomorphic to the inferential relations between dog, animal, pet
and so on.[3]
In the face of
this problem, the natural move is to allow that some concepts have their
reference fixed by something other than their inferential role. But this move will then require some
explanation of representation than goes beyond purely inferential role semantics. It remains possible that inferential role
semantics alone can explain the content of some concepts, once the
contents of others have been explained in some different way. However, I shall not pursue this possibility
here, since it leaves inferential role semantics with only a derivative part in
explaining reference, and moreover still facing the problem of conceptual
holism.
3 Causal Theories
Another family of naturalist theories of meaning aims to explain the
representational content of mental states in terms of the conditions that cause
those states, and which those states therefore indicate (Stampe, 1977, Dretske,
1981, 1988, Fodor 1990). At its
simplest, such a theory might start by equating the content of any belief-like
mental state B with that condition C which is causally responsible for all
tokens of B.
This simple theory
is clearly too crude, however, since it lacks the resources to explain misrepresentation. Misrepresentation by a belief-like state
occurs when the state is tokened, but its truth condition does not obtain. However, if the state’s truth condition is
simply the range of circumstances that cause the state to be tokened, then it
is unclear how the state can be tokened and yet its truth condition not
obtain.
To make the problem
clear, take a state that intuitively represents the presence of a snake. Such a state will often be caused, not by
real snakes, but also by glimpses of slithery animals, toy snakes, and so on. The problem for the simple causal theory is
that it has no obvious way of excluding these misleading extra causes from this
state’s truth condition. So the causal
theory seems to end up implying, absurdly, that all tokenings of this
belief-like state are true.
Fred Dretske
(1981) develops a version of indicator semantics that is designed to account
for misrepresentation. He argues that
the truth condition of a belief-like state B should be identified specifically
with the causes of tokens of B that occur during ‘the learning period’, that
is, during the period when the disposition to produce tokens of B is reinforced
by experience. This then leaves room for
tokens of B produced outside the learning period to misrepresent, since they
might or might not be due to the same causes that operated during the learning
period.
While Dretske’s
theory does leave room for misrepresentation, it faces other difficulties. For one thing, it presupposes a sharp
distinction between the learning period (when misrepresentation is impossible)
and subsequent tokenings of B (which can misrepresent), even though there seems
no principled basis in psychological learning theory for such a
demarcation. Another problem is that
there seems no good reason why the causes that do operate during the learning
period should automatically be included in B’s truth condition: for example, a child might learn to represent
snakes by observing toy snakes or pictures of snakes, yet toy snakes and pictures
of snakes are not part of the truth condition of snake.[4]
Jerry Fodor (1990)
defends a different version of indicator semantics. His basic idea is to discriminate fundamental
from derivative causes of B, and to equate truth conditions with the
fundamental causes. By way of example, note
that the belief there’s a cow can be caused by cows, but also by horses
at some distance. However, the
relationship between horses and this belief is only derivative, argues Fodor,
in that horses wouldn’t cause this belief if cows didn’t, whereas cows would
still cause this belief even if horses didn’t.
According to Fodor’s asymmetric dependence theory, B represents C
just in case (i) C causes Bs and (ii) for any other D that causes Bs, D
wouldn’t cause Bs if C didn’t cause Bs, while C would still cause B even if D
didn’t. On this account, then, the
belief that there’s a cow represents cows but not horses, because of the
asymmetric way this beliefs depends on the cows and horses respectively.
The basic worry
about this theory is that it seems in danger of implicitly supposing what it is
supposed to explain. Who says that cows
would still cause the mental state that actually has the content there’s a
cow, even if horses didn’t? After
all, it is pretty inevitable that people are always going to mistake a few
horses for cows. So if some state were
never caused by horses, then surely it would follow that it couldn’t mean there’s
a cow. However, if this is right,
then Fodor’s counterfactuals will fail to discriminate cows from horses as the
referent of there’s a cow, since neither horses nor cows would cause
this state if the other didn’t. In the
light of this objection, it looks as if Fodor must implicitly be holding fixed
the actual content of the mental state when he insists that cows would still
cause this state, even if horses didn’t.
But this would be illegitimate, in a context where the counterfactuals
are supposed to provide a metaphysical reduction of representational content.
4 Success
Semantics
All causal
indicator theories share one important feature.
They focus on the conditions that give rise to belief-like
representations, aiming to equate truth-conditional content with some
distinguished subset of these ‘input’ conditions. A different family of theories does things
the other way around. Instead of
starting with the conditions that give rise to representations, they focus on
the consequences of representations.
Such ‘output-orientated’ theories include success semantics and
teleosemantics. I shall discuss success
semantics in this section and teleosemantics in the following sections.
According to
success-semantics, the truth condition of any belief is that circumstance which
will ensure the satisfaction of whichever desire combines with the belief to
prompt action. (Ramsey, 1927, Appiah,
1986, Whyte, 1990, Dokic and Engel, 2002.)
More intuitively,
the idea is that beliefs are dispositions to behaviour—what makes it the case
that you believe p is that you behave in a way that will satisfy your desires if
p. For example, you believe that there
is beer in the fridge if you go to the fridge when you want a beer.
Success semantics
has no difficulty accommodating misrepresentation. Because it analyses truth conditions in terms
of results, rather than causes, it carries no implication that beliefs will
generally tend to be true. The content
of a belief is fixed by the behaviour it generates, not by the causes that give
rise to it. As long as it makes me go
the fridge, my state will have the content that there is beer there, even if
this state is characteristically caused when there is no beer in the
fridge. Success semantics thus creates
ample room for beliefs to be false, even typically false.
One obvious
problem facing success semantics is that many beliefs will only combine with
desires to generate behaviour if they are conjoined with yet further
beliefs. (Consider, for example, the
belief that the sun has nine planets.)
To deal with this, success semantics needs a more complicated
formulation: the truth condition of any
belief is that circumstance which will ensure the satisfaction of whichever
desire it combines with to prompt action, on the assumption that any other
beliefs involved in generating that action are true.
However, as it
stands this is obviously inadequate as a reductive account of truth-conditional
content, since the last clause assumes the notion of truth. The most promising way for success semantics
to overcome this difficulty is to regard the connection between truth
conditions and desire satisfaction as being imposed simultaneously on all the
beliefs in a thinker’s repertoire. We
get the truth condition for all these beliefs by solving a set of simultaneous
equations, so to speak. The ‘equations’
are the assumptions the truth condition of each belief guarantees desire
satisfaction, if all other relevant beliefs are true. The ‘solution’ is then a collective
assignment of truth conditions that satisfies all those equations.
There is another
obvious objection to success semantics.
In explaining truth conditions, it assumes the notion of desire
satisfaction. But desire satisfaction is
itself a representational notion, and so cannot be taken for granted by a
reductive theory of representation.
The natural
response to this difficulty is to find some independent account of desire
satisfaction (Whyte, 1991.) One possibility
is to equate satisfaction conditions for desires with those circumstances that
typically extinguish the desire—my desire is a desire for beer because
it is beer that makes that desire go away.
An alternative is to equate satisfaction conditions with those results
that are reinforcing—that is, which make it more likely that the behaviour
prompted by the desire will be repeated next time the desire is activated.
However, it is not
clear that either of these suggestions is fully satisfactory. The equation of satisfaction conditions with
extinguishing circumstances has difficulties with desires that are fuelled by
their own satisfaction (salted peanuts) or quenched by their non-satisfaction
(sour grapes). Again, the explanation of
satisfaction in terms of reinforcement seems to rule out the possibility of
desire content where there is no reinforcement learning, even though this would
seem a real possibility, both for primitive creatures with limited behavioural
flexibility and for humans in respect of their more abstract desires.
5 Teleosemantics
One way of
understanding teleosemantics is as a combination of success semantics with a teleological
account of desire satisfaction. (Cf.
Papineau, 1993, ch 3.) So conceived,
teleosemantics embraces the connection between truth conditions and desire
satisfaction articulated by success semantics, and then deals with the problem
of explaining desire satisfaction by equating satisfactions conditions with the
biological functions of desires.
(The notion of biological ‘function’ invoked here is in turn explained
in aetiological-selectional terms: the biological
functions of desires are those results in virtue of which the desires have been
favoured by past processes of natural selection.)
This approach to
teleosemantics is ‘top-down’, in that it takes a realistic attitude to human
belief-desire psychology, and then seeks a naturalistic account of
representation for the human beliefs and desires it is thus committed to. Ruth Millikan has developed a more generalized
‘bottom-up’ version of teleosemantics, aimed in the first instance at representation
in organisms far simpler than human beings (1984, 1993). Millikan starts by distinguishing mechanisms
that produce mental representations from those that consume them.
The producing mechanisms are
paradigmatically the sensory processes that give rise to cognitive
representations. The consumer mechanisms
are those that use these representations to direct behaviour in pursuit of some
biological end. Millkan then considers
the biological functions of mental representations. Biological functions are in the first
instance always a matter of effects. So the
function of a mental representation must lie in the way it contributes to the
biological end of the mechanism that consumes it. More specifically, its function will be to
enable the consumer mechanism to achieve its end by gearing behaviour to
circumstances. Given this, argues
Millikan, we can think of the representation’s truth condition as the
circumstance that enables it to fulfil this function—that is, as the
circumstance in which the behaviour it prompts is designed to produce the
consumer mechanism’s end.
Millikan’s version of
teleosemantics coincides with the version that builds on success semantics if
we equate the consumer mechanism for a belief with the decision-making process
that uses that belief to select behaviour that will satisfy currently active
desires. Given this, the association of
a success condition with a belief can be viewed as one example of the way
Millikan’s analysis fixes the content of any belief-like representation. At the same time, Millikan’s version of
teleosemantics is far more general than the success-semantics alternative, in
that it can also deal with representation in creatures who lack the cognitive
complexity of full belief-desire psychology.
Millikan’s bottom-up strategy
has the obvious advantage of more general applicability, and moreover avoids
the danger that everyday belief-psychology may offer a misleading picture of
actual human cognitive structure. On the
other hand, a full account of mental representation will need to cover human
cognition too, and the top-down approach via success semantics offers one
possible account of this. In the end,
perhaps the two approaches are best thought of as complementary rather than
competing.
In what follows I shall
ignore the differences between these versions of teleosemantics and concentrate
on issues that arise for both. The next subsection
will focus on the output-orientation of teleosemantics, while the final subsection
will be concerned with teleosemantics’ commitment to selectional functions.
5.1 Teleosemantics and
Outputs
One strength of
teleosemantics is that it inherits the ability of success semantics to deal
with misrepresentation. Since teleosemantics
is also output-based, it coincides with success semantics in imposing no
requirements on how representations are caused, nor on the reliability of the
mechanisms which produce them. A
representation can have the content that p, in the sense that resulting
behaviour will be successful if p, even if its producing mechanisms are highly
prone to give rise to the representation when not-p.
Moreover, now that we are
thinking of representation in a specifically biological perspective, this
divergence between truth-conditional content and typical causes is no longer
merely an abstract possibility. Consider
a small mammal which can form a representation which will lead it to behave in
a way appropriate to an eagle being overhead.
According to teleosemantics, this representation will have the content
‘eagle overhead’, since its purpose is to prompt behaviour which will be
advantageous specifically in that circumstance.
However, given that the relative biological costs of false positives and
negatives in this context, we can expect that the mechanisms which produce this
representation will err generously on the side of caution, and frequently
trigger the representation in circumstances where no eagle is in fact overhead.
Not everybody regards
this input-independence as an obvious virtue in teleosemantics. If the small mammal’s representation is
triggered by any moving shadow, say, would it not be better interpret its
content as ‘moving shadow’ rather than ‘eagle overhead’?
This reaction is
bolstered by the following well-known thought-experiment due to Paul Pietroski
(1992). The kimu are simple creatures,
with very limited sensory abilities, whose only enemies are the snorf, who hunt
them every day at dawn. A mutation
endows one of the kimu with a disposition to sense and approach red
things. This disposition is a biological
advantage to its possessors, since it leads them to climb a nearby hill every
dawn, the better to observe the red sunrise, and means that they thereby avoid
the marauding snorf, who do not climb hills.
As a result, the disposition spreads through the kimu population.
Now, consider the state
a kimu gets into when it is stimulated by something red. It seems natural to credit this state with
the content red. But an
output-based teleosemantics argues differently.
Nothing good happens to the kimu just because they approach something
red. Most of their red-approaching
behaviour is just a waste of time. It is
only when this behaviour takes them away from the dangerous snorf that it
yields any biological advantage. So an
output-based teleosemantics will deem the state in question to represent snorf-free,
or predator-free, or some such. This
strikes many as strongly counter-intuitive.
After all, by hypothesis the kimu’s senses are tracking the presence or
absence or redness, not the presence or absence of snorfs.
Still, advocates
of teleosemantics can respond that these intuitions depend on reading more into
Pietroski’s scenario than is justified by his description. Pietroski says that the kimu evolve some
state that is triggered by redness and which has the advantage of keeping them
away from the snorf. Given this
specification, it is natural to think of the kimu as having some
general-purpose visual system which gathers items of visual information
which informs an open-ended range of behavioural projects directed at different
possible ends (such as avoiding blood, or finding post-boxes, or indeed wanting
to see red things). However, this extra
structure in fact takes us significantly beyond what Pietroski’s description
actually requires, and it is open to teleosemanticists to argue that their
theory is quite able to explain why an organism with all this extra structure
would be representing redness rather than snorf-freeness: if the organism’s visual states inform a range
of different behaviours directed at different ends, then the content of any
such state needs to be fixed as some condition that assists in the achievement
of all those ends, and this may well come out as redness. On the other hand, if we do stick to a
minimal understanding of the snorf, as having only a special-purpose visual
sensitivity that brings no advantage except snorf-avoidance, then it’s not so clear
that there is anything wrong with the output-based reading of their states as
representing snorf-freeness:
after all, if these states never do anything except trigger simple
avoidance behaviour, it seems natural enough to read them as representing the
danger they are designed to avoid.
It might seem unclear
why teleosemantics is forced to focus exclusively on output conditions and
ignore input conditions in explaining content.
What would be wrong with a hybrid input-output theory which starts with
the relationship between input conditions (red surfaces) and representations
(the kimu state), and then says that such a correlation constitutes representation
if it is has the function of guiding behaviour in pursuit of biological ends? Something like this approach has been
explored by a number of writers (Neander, 1995, Dretske, 1988, Millikan, 2004). However, it is not clear that this leads to a
substantial alternative to a purely output-based teleosemantics. Remember that many different input
circumstances will be correlated to a greater or lesser degree with any given representation-type—including
all those causes that systematically give rise to misrepresentations. So, even if we start with input conditions,
we still face the task of explaining what picks out the representation’s genuine
content from all the other potentially informational input-representation
correlations. And then we will be back where we started, if the only answer is
that content corresponds to that correlation which is a matter of biological
design. For the biological function of
representation is to guide behaviour in pursuit of biological ends, and so an appeal
to biological design can do no other than pick out as content that circumstance
required for the organism’s behaviour to yield biological success (which for
the kimu will once more be snorf-freeness rather than redness).[5]
5.2
Teleosemantics and Selection
According to
teleosemantics, representational content depends on biological design, and
biological design requires a history of natural selection. This prompts an obvious query. What about creatures who have no such
history? Will they not be able to
represent?
This worry is
normally pressed with the help of the ‘swampman’ thought experiment (Davidson,
1987). Suppose that lighting strikes a
steamy marsh deep in the tropical jungle, and that by miraculous conincidence a
perfect molecule-for-molecule replica of a human being assembles itself out of
the organic materials available in the swamp.
By hypothesis, this ‘swampman’ will lack any history of natural
selection, and so, according to teleosemantics, will not be possessed of any
representational powers. Yet intuitively
it seems that swampman will be perfectly capable of at least some forms of
mental representation. After all, it will
be physically just like a normal human, so will be equally capable of visually
registering its surroundings and making appropriate behavioural responses. So it
looks as if teleosemantics has gone wrong somewhere, if it denies that swampman
has any representational capacities.
The standard
teleosemantic response to this difficulty is to bite the bullet and maintain
that swampmen will indeed be incapable of representation. Maybe everyday intuition says that swampmen
can represent. But a good theoretical
account should be allowed to overturn a few everyday intutions. Just as our modern concept of fish excludes
whales, despite any naïve intuitions to the contrary, so should a developed
concept of representation be allowed to exclude swampmen. According to this line of thought, then, we
should replace our naïve concept of representation by the theoretically more
powerful selection-based notion, even at the cost of overturning intuitions
about swampmen. (Cf. Millikan, 1996, Neander, 1996, Papineau,
1996.)
However, there is room
for an alternative and more irenic defence of teleosemantics against swampman
worries. The alternative strategy is to
leave the concept of representation as it is, and focus instead of the status
of teleosematics as an a posteriori reduction of representational
facts—that is, as a scientific theory that reveals the selectional nature of
representation, just as chemistry reveals the nature of water to be H20. From this perspective, it is no argument
against teleosemantics that representationally competent swampmen are
consistent with our everyday concept of representation; you may as well oppose against modern
chemistry on the grounds that XYZ-composed water is consistent with our
everyday concept of water. The fact that
swampmen with representations can be imagined does nothing to undermine
the central teleosemantic claim that in the actual world
representational facts consist of selectional facts. Of course, if there were plenty of actual
swampmen, then things would be different, for they would then provide concrete
evidence that teleosemantics is false.
But as long as swampmen remain merely imaginary, they are no more
relevant to teleosemantics than imaginary molecular make-ups are relevant to
chemistry. (Papineau, 2001.)
Let me conclude this
discussion of teleosemantics by addressing one further worry about the appeal
to selection. Teleosemantics takes all
representation to depend on histories of natural selection. The most familiar kind of such natural
selection is the intergenerational selection of genes. However, it is surely unlikely that all
representation can be explained in terms of such genetic selection. After all, most human beliefs and desires are
products of ontogeny rather than phylogeny.
No genes have been selected specifically to foster those specific
beliefs or desires.
Fortunately for the teleosemantic project, the
ascription of a selectional function to some trait does not always require that
specific genes have been selected because they give rise to those traits. There are ways in which biological items can
have aetiological-selectional functions even though they have no specific
genetic basis. In particular, there are
two theoretical possibilities that often go unnoticed in this context. The first, emphasized by Millikan, appeals to
a many-layered account of functions. The second appeals to non-genetic
selection. Together these greatly expand
the range of items that possess aetiological-selectional functions.
Multi-layered
functions first. Millikan’s notes that one
kind of function is a relational function, which is a function to do
something only when bearing a certain relation to something else. The chameleon's skin-colour mechanism has the
relational function of making the chameleon's skin-colour match that of its
environment, whatever that colour may be.
Given a specific colour to adapt to, this mechanism then generates
traits with derived functions. When
the chameleon is sitting on a brown plant, its skin colour has the derived
function of matching it to the brown environment. Note that this brown skin may have never been
produced before, but even so will have a derived function, in virtue of the
fact that the skin-colour mechanism has been selected to produce whatever
colour will match the background.
This analysis in
terms of multi-layered functions can be applied to novel representations within
compositional syntactical system.
Consider the famous dance of the bees, which acts as a signal to other
bees, ‘telling’ them where to go to find nectar. A particular dance will be adapted to the current
location of the nectar, and so will have a derived function. Again, the dance that indicates this specific
direction may never have occurred before.
Rather, it owes its functionality to the function of a system
that has yielded a reproductive advantage in the past. Analogously, we can expect that many features
of human cognition can be viewed as having biological functions, not because
they themselves have been selected for, but because they are products of a system
that has been so selected.
The other non-genetic source of
aetiological-selectional functions is non-genetic selection. There are two possibilities worth mentioning
in this context. One is selection-based
learning. This doesn’t involve
the differential reproduction of organisms over generations, but the
differential reproduction of cognitive or behavioural items themselves during
the development of a given individual.
Such ontogenetic selection takes place, for example, when behaviour is
moulded by experience during learning.
In such cases we can think of the items selected as having the function
of producing those effects in virtue of which they were favoured by the
learning mechanism.
The other kind of non-genetic selection relevant to
teleosemantics is non-genetic intergenerational selection. Many traits are passed from parents to
children by channels other than the sexually transmission of genetic
material: these traits will include the
possession of parasites, the products of imprinting mechanisms, and the many
cognitive and behavioural traits acquired from parents via social
learning. A number of biological
theorists are currently interested in the way in which such non-genetically
inherited traits can be naturally selected through the normal Darwinian process
of differential reproduction of organisms (Jablonka and Lamb, 1999, Mameli,
2004). Non-genetically inherited traits
that become prevalent in this way will have functions, namely, the effects
which favoured their possessors. It seems
highly possible, though this is an area that has yet to be properly explored,
that functions of this kind could do much to explain the contents of
sophisticated mental representations.
After all, it seems a natural enough thought that certain non-genetically
inherited ways of thinking are an advantage their possessors because they make
them sensitive to certain features of the environment. On the other hand, it remains an open
question how many features of human thought are in fact due to differential reproduction
of offspring resulting from such advantages.
It should be said that that there is as yet little
detailed work showing how teleosemantics might analyse sophisticated human
modes of cognition by appealing to functions other than those deriving directly
from the selection of genes. True,
Millikan (1984) has indicated how her notion of an adapted proper function can
be used to account for the representational contents of elements in complex
representational systems. And Dretske (1988)
has focused on selection in learning as one means by which to explain how
cognitive states can be teleosemantically targeted on specific contents. Still, much remains to be done in applying
teleosemantics to specifically human modes of cognition.
Perhaps this is inevitable. Detailed analyses of representational powers
in terms of aetiological functions must rest on an adequate empirical knowledge
of the cognitive mechanisms involved.
There is no question of identifying the functions of cognitive items if
we don’t know what kinds of mechanisms process these items and how those
mechanisms develop in individuals. From
this perspective, the teleosemantic project is not so much a theory of content
for sophisticated human representation, but a methodology which promises to
explain content piecemeal, in the wake of empirical discoveries about human
cognitive architecture.
References
Block, N. 1986. ‘Advertisement for a semantics for
psychology.’
Cummins, R. 1991. Meaning and Mental Representation.
Churchland, P. 1989. A Neurocomputational
Perspective: The Nature of Mind and the Structure of Science.
Davidson, D. 1987. ‘Knowing ones own mind.’ Proceedings
and Addresses of the American Philosophical Association 60.
Dokic, J. and Engel, P. 2002. Frank Ramsey: Truth
and Success.
Dretske, F. 1981. Knowledge and the Flow of
Information.
Dretske, F. 1988. Explaining Behaviour.
Fodor, J. 1974. ‘Special
Sciences: Or, The Disunity of Science as a Working Hypothesis.’ Synthese
28.
Fodor, J. 1990. A Theory of Content and
Other Essays.
Fodor, J. and Lepore, E. 1992. Holism: A Shoppers’ Guide.
Harman, G. 1982. ‘Conceptual role semantics.’ Notre
Dame Journal of Formal Logic 23.
Harman, G. 1987. ‘(Nonsolipsistic) conceptual role
semantics’ in LePore, E. (ed.) Semantics
of natural language.
Horwich, P. 1990. Truth.
Jablonka, E. and Lamb, M. 1999. Epigenetic
Inheritance and Evolution.
Mameli, M. 2004, ‘Nongenetic
Selection and Nongenetic
Inheritance’, British
Journal for the Philosophy of Science 55.
Millikan, R. 1984. Language
Thought and Other Biological Categories.
Millikan, R. 1993. White
Queen Psychology and Other Essays for
Millikan, R. 1996. ‘On Swampkinds.’ Mind and Language 11.
Millikan, R. 2004. Varieties
of Meaning.
Neander, K. 1995.
‘Malfunctioning and Misrepresenting.’ Philosophical Studies 79.
Neander, K. 1996. ‘Swampman Meets Swampcow.’ Mind and Language 11.
Papineau, D. 1993. Philosophical
Naturalism.
Papineau, D. 1996. ‘Doubtful Intuitions.’ Mind and Language 11.
Papineau, D. 1998. "Teleosemantics and Indeterminacy.’
Australasian Journal
of Philosophy 76.
Papineau, D. 2001. ‘The
Status of Teleosemantics, or How to Stop Worrying about Swampman.’ Australasian Journal of Philosophy 79.
Papineau, D. 2003. ‘Is Representation Rife?’ Ratio
16.
Peacocke, C. 1992. A Study of Concepts.
Pietrowski, P. 1992.
‘Intentional and Teleological Error.’ Pacific Philosophical Quarterly 73
Ramsey, F. 1927. Facts
and Propositions. Aristotelian Society Supplementary Volume 7.
Stampe, D. 1977. ‘Toward
a Causal Theory of Linguistic Representation’ in P. A. French, T. E. Uehling,
Jr., and H. K. Wettstein (eds) Midwest
Studies in Philosophy: Studies in the Philosophy of Language 2.
Quine, W. 1951. ‘Two Dogmas of Empiricism.’ Philosophical
Review 60.
Whyte, J. 1990. ‘Success Semantics.’ Analysis
50.
Whyte, J. 1991. ‘The
[1] The view that mental representation is more basic than public language allows that creatures without any public language might nevertheless have mental representations. But at the same time it leaves open the possibility that many human mental representations may be developmentally or even constitutively dependent on surrounding linguistic practices.
[2] This
eliminativist position is defended in Churchland, 1989. An alternative to outright eliminativism
about representation is offered by ‘minimalist’ or ‘disquotationalist’ views
which hold that the truth predicate does not refer to any substantial property,
but is rather a device for endorsing claims without asserting them (cf.
Horwich, 1990).
[3] Cummins (1992) is prepared to embrace such
a referentially promiscuous notion of representation, at least for the purposes
of cognitive science.
[4] Dretske (1988) adds a teleological
component to his causal theory of representation, but difficulties relating to
the learning period remain.
[5] Teleosemantics is often
charged with an inability to ascribe determinate contents to cognitive states
using only considerations of biological design.
(Cf. Fodor, 1990.) An initial
response to this charge is that only those input-representation correlations
that are used in pursuit of biological ends qualify as
representational. The adequacy of this
response depends in large part on how far the relevant notion of biological end
can be rendered unequivocal. (Neander,
1995, Papineau, 1988, 2003.)