Millikanian Teleosemantics

Baseley 1

Millikanian Teleosemantics

Logan Baseley

Phil 5478

16 November 2014

University of Missouri-St Louis

Baseley 2

Department of Philosophy

Logan Baseley

PHIL 5478

16 November 2014

Term Paper

Millikanian Teleosemantics

Abstract

A belief that the sun is rising is an example of a mental

state. Mental states such as beliefs are about things; the

belief that the sun is rising is about the sun rising. That the

sun is rising is the content of the belief. If one can talk,

perhaps only metaphorically, of mental states giving attention,

then mental states give attention to things, namely, states of

affairs. I take this to be a fairly uncontroversial position.

Justin Garson evinces the same confidence. He says,

unequivocally, that “thoughts are about things.”1 (Garson

1 Garson, Justin. The Biological Mind 137

Baseley 3

actually repeats this line, twice, in the same paragraph. How’s

that for confidence!) That being the case, I do not attempt to

defend representationalism per se. Such a defense is well beyond

the scope of this paper. My objective is much narrower than

that. I analyze Ruth Millikan’s teleosemantic account of mental

representation and describe its scope, highlights, and goals.

Then I take a critical position. I entertain the idea that

Millikan’s position cannot give an account of accurate

representation, but only one that tends toward biological

fitness. I consider this a strong objection. Moving forward, I

anticipate an objection from the Millikanian camp and give ground

to Millikan-style teleosemantics. Finally, I conclude that

Millikan-style teleosemantics can stand up to my objection.

Introduction

A tradition of naturalizing semantics began in the late

1970s. Many attempts were and still are being made to bring

naturalistic-physical explanations to bear on minds and,

specifically, to the question of how minds acquire content.2

2 Adams, Fred. “Thoughts and Their Contents: Naturalized Semantics.” In Philosophy of Mind 143.

Baseley 4

This is an interesting question; it is no surprise that it takes

center stage in the philosophy of mind. Indeed, it is certainly

an interesting question how minds, thought by those in the

naturalism camp to be “natural physical objects,”3 could have

developed intentional properties. In the mid-1980s, with the

works of Ruth Millikan, Karen Neander, and David Papineau

(Language, Thought, and Other Biological Categories and “Representation and

Explanation,” respectively) teleosemantics, a theory of mental

content that attempts to address the question of content and

intentionality of minds, was born.

Teleosemantics, a flag bearer in the naturalizing content

camp, attempts to provide an explanation for the intentionality,

or aboutness, of mental states, e.g., thoughts and beliefs. In

this paper I focus, specifically, on Millikan’s version of

teleosemantics. Hers differs from other teleosemantic

explanations of representation in, as I see it, one important

way. To wit, Millikan draws a distinction between two

representational systems: producer systems and consumer systems,

each of which is located within a single perceiver or perceptual

3 Ibid.

Baseley 5

system. Millikan, like all other Teleosemanticists, brings to

bear the notion of biological function in explaining the

intentional character of mental states. Whereas the biological

function of the testes is to reproduce (in conjunction with other

organs), the biological function of mental states is to represent

content.4

I begin, in section one, by providing an in-depth

description of Millikan’s teleosemantics, which I understand as a

bio-functional explanation of perceivers’5 representational

capacities. I then take a critical position. My objective in

section two is to tentatively show that Millikan’s account of

representation-producing states does not necessarily entail

accurate or veridical representations of worldly affairs;

representation-producing states must produce representations that

tend toward an organism’s fitness, but these are not necessarily

accurate representations. I take this to be an indictment against

Millikanian teleosemantics: her theory can give us an account of

representation that tends toward an organism’s biological

4 Garson, Justin. The Biological Mind. 140.5 (Perceivers like us.)

Baseley 6

success, but not an account of veridical representation. In the

next section, I anticipate a reply from the Millikanian camp—that

accurate or veridical representation need not tend toward fitness—

and give ground to Millikan-style teleosemantics. Finally, I

conclude that Millikan-style teleosemantics can stand up to my

objection. I conclude that Millikan-style teleosemantics is on

solid footing.

Millikan-Style Teleosemantics

Teleosemantic accounts of mental representation attempt to

explain perceivers’ ability to represent states of the world on

the basis of biological functions. Teleosemantics uses the

notion of function to explain representation. Teleosemanticists

are, among other things, interested in describing how mental

intentionality arises naturally. Before moving on, we need to

unpack the concept of biological function. Take the heart and

its pumping blood—a classic example that illustrates the notion

of biological function. The heart pumps blood; in pumping blood,

it makes a thumping sound. But it is the pumping of blood, and

not the thumping sound, that is the heart’s function; this is

Baseley 7

because pumping blood, and not making sound, led—evolutionarily—

to the system’s fitness.6 Lycan puts it quite clearly, saying “a

heart’s function is to pump blood because hearts’ pumping blood

in the past has given them a selection advantage and so led to

the survival of more animals with hearts.”7 The fact that

hearts’ pumping blood in the past led to survival plays a

significant explanatory role in Millikan’s account of biological

functions.

Following the paradigmatic case of the heart, we can begin

to see how other capacities and traits can be explained in terms

of functions. Take the intentional character of mental states.

It’s been noted that beliefs are about something. What accounts

for the aboutness, or intentionality, of mental states? Why does

that trait exist in a system of mental states? Ruth Millikan,8

perhaps the most prominent teleosemanticist, argues that

teleology must play a crucial role in explaining the

intentionality of mental states. On her account, teleology is

given explanatory pride of place in analyzing the intentionality

6 Wright, Larry. “Functions” Philosophical Review 82: 139-168.7 Lycan, William. “The Mind-Body Problem.” 55. In The Philosophy of Mind.8 Millikan, Ruth. Language, Thought, and Other Biological Categories.

Baseley 8

of mental states. In order to answer the question of why

intentionality exists in a system of mental states, we must first

analyze more closely Millikan’s conception of representation.

Millikan’s view differs from other teleosemantic views in

myriad ways, but perhaps its most unique characteristic is its

distinction between the mechanisms that produce mental

representations from those that consume mental representations.9

There is a representational function as a whole, at a composite

level; and there are two “sub-functions,”10 the producer-function

and the consumer-function. In terms that are easy to understand,

let’s take Millikan’s own example of beavers’ splashing their

tails. One beaver alerts other beavers to the presence of danger

by splashing its tail on the surface of water. The splashing of

the tail tells, or represents to, the other beavers that there is

danger in the environment, and the other beavers dip into the

water to avoid the danger. The splashing of the beaver’s tail

produces a representation; the other beavers consume the

9 Macdonald, G, and Papineau, D. “Prospects and Problems for Teleosemantics,” in Teleosemantics 4.10 Shea, Nicholas. “Millikan’s Contribution to Materialist Philosophy of Mind,” 6.

Baseley 9

representation. The representation that the beavers consume

guides their behavior in ways that relate to their survival.

Now, of course, we are interested in representations within

one cognitive system, and not just representational states of

affairs between two (or more) distinct organisms (or cognitive

systems). How does the picture of the producer and consumer

beavers, for instance, play into a story about internal

representations? Papineau and Macdonald describe Millikan’s

account of this well and loyally, saying “The producing

mechanisms will be the sensory and other cerebral mechanisms that

give rise to cognitive representations.”11 The consuming

mechanisms are those that “use these representations to direct

behavior in pursuit of some biological end.”12 Here, we have a

picture similar to the beaver example, but this picture portrays

the two sub-functions, producer and consumer, operating within a

more-obviously unified system, namely, the cognitive system. One

sub-function produces mental representations while the other sub-

function consumes them in order to reach some end, e.g., danger-

11 Macdonald, G. and Papineau, D., “Prospects and Problems for Teleosemantics,” in Teleosemantics 4.12 Ibid.

Baseley 10

avoidance or food-acquisition. The representations consumed by

the consumer sub-function guide an organism’s behavior toward

some biological end, e.g., survival. This is a rather brief

sketch of Millikan’s overall portrait. Of course, more goes into

her account of the relation between producer- and consumer-

functions in order to arrive at a nuanced account of mental

representation. But that is a matter of how. Details as to the

how aside, much of Millikan’s efforts are directed towards the

why, viz., why it is that perceivers like us have mental

representations—why representations are produced in the first

place. Below, I formalize Millikan’s account of the ‘why’ in

relation to functions:

PI. If a system has a function, then it has a use.

PII. Something has a use only if there is a system capable of

using it.

PIII. The function of a mental state is to produce

representations.

C1. Thus, the representations that are produced have a use.

C2. Thus, there is a system capable of using the representations

that are produced.13

13 Millikan, Ruth. “Useless Content,” in Teleosemantics, 100.

Baseley 11

Millikan wants to show that the function of a mechanism is

understood in relation to the function’s purpose—how the function

contributes to the organism as a whole. Further, she wants to

say that the function of a mental state is to represent the world

as being such-and-such way. In order to say this, though, she

must explain such a representation’s use and, indeed, the use of

the representational system itself. She describes the

representational system’s use by describing its relation to and

cooperation with the consuming system, i.e., that biological

system whose use of the representations informs its behavior

(behavior that is aimed at some survival-related end). What

follows from this, then, is that the producing system needs to

represent the world in specific ways that allow the consuming

system to respond appropriately, i.e., in a way that leads to the

fulfillment of ends.

In what way must the producing system represent the world?

Must the producing system always represent the world veridically,

i.e., matching reality; or can the producing system produce

representations that do not represent veridically but, for

instance, in ways that lead the consuming system to err on the

Baseley 12

side of caution? To come close to answering this question, it

might help to look at a specific example of internal

representation and how its representation-conditions tend towards

effective goal-acquiring behavior. (By ‘representation

conditions’, I mean those conditions that directly affect a given

behavior’s efficiency.)

Representation Conditions for Millikan

In what way must the representation system represent the

world? The producing system cooperates with the consuming

system. The consuming system receives the representations

produced by the producing system. The nature of this cooperation

involves the producing system producing representations that

allow the consuming system to produce behavior, effectively,

towards some end (acquiring food, avoiding predators, etc.) The

producing system must, then, represent the world in ways that

allow the consuming system, and the organism as a whole, to

achieve its goals; otherwise, the organism as a whole would fail

to exhibit behavior necessary for survival. We could not

describe the producing and consuming systems in terms of

Baseley 13

functions, given this failure to survive, since (as has been

said) without a use, a system has no function (this is Premise II

from above).

Thus, the producing system must produce representations in a

way that causes the consuming system’s reception of the

representations to lead to appropriate behavior. Just to be

clear: if the behavior of the consuming system is going to bring

about the effects it was designed to bring about—through natural

selection—then the producing system must cooperate with the

consuming system in a particular way. The producing system must

produce representations that are going to allow the consuming

system to bring about its designed-for effects. This is all well

and good, and one may argue against Millikan on, say, the fine

grain details of the physiological route between produced-

representations and consumed-representations. Indeed, what may

need to be addressed, by Millikan, is whether, for instance, the

Rana pipiens’ (or leopard frog) mode of producing and consuming

representations is similar to the way in which our producing and

consuming systems operate. Whereas for the leopard frog

Baseley 14

“information”14 travels from the retina, to the optic nerve, to

the optic tectum, and finally to the motor system, is this

picture of cooperation between producer and consumer similar to

our own producer-consumer mechanism?15 This is, indeed, an

interesting question—but one beyond the scope of this paper.

Presently, I am less interested in the ‘physiological route’

question described above, and more interested in veridicality.16

As I said above, the producing system must produce

representations that allow the consuming system to bring about

its designed-for effects. As I see it, the producing system can

do this in two ways. First, the producing system can produce

representations, which the consuming system will consume, that

are wholly veridical. That is, the representations will totally

coincide with reality; every representation will mirror the thing

in the world being just as it is. This option seems unlikely; in

our own experiences, we seem to sometimes fail to represent

veridically. Second, the producing system can produce

representations, which the consuming system will consume, that 14 Garson, Justin. The Biological Mind. 138.15 Ibid.16 As far as I can tell, Millikan does not address the physiological question directly. Neander has something to say about it. But, like I said, I have found no response by Millikan herself.

Baseley 15

are not wholly veridical but tend toward the fitness of the

organism all the same. The representations will not totally

mirror reality but will, rather, be consumed in such a way as to

enable the consuming system to exhibit behavior that will produce

the effects it was designed to produce. The second account, of

course, presupposes that effects can be produced by the consuming

system without veridical representations; I address this

presupposition below.

Let us consider, first, whether the representations of a

producing system need, on Millikan’s view, to be wholly

veridical. Intuitively, it seems plausible that they need not

be; we can think of representations of things in the world that

are not wholly veridical but that nevertheless contribute to an

organism’s biological success. Consider this twist on Paul

Pietroski’s well-known thought experiment concerning kimus and

snorfs.17 Bogs are simple creatures with simple survival

requirements. Bogs need to eat very little; in fact, they can

skip meals and go for days without food. Bogs are also quite

timid. The slightest shadow makes bogs flee whatever area

17 Pietroski, Paul. “Intentional and Teleological Error,” Pacific Philosophical Quarterly, 73: 267-82.

Baseley 16

they’re in. Bogs almost always sacrifice food acquisition for

danger avoidance. The cautious nature of bogs has given them a

selection advantage. Now, imagine that the producing system of

bogs—the cerebral mechanism that gives rise to cognitive

representations—constantly produces representations of shadows.

The representations of shadows are then consumed, by the

consuming mechanism, and used by the mechanism to direct behavior

that is aimed at some end—in this case, escaping a perceived

threat. The consuming system interprets the shadows as ‘shadow-

dangers’. What we have in this case is a mismatch between the

meaning of what is represented, a shadow-danger, and what is, in

reality, really there—just a shadow. Yet, despite this mismatch,

the representation mechanism is doing what it is supposed to be

doing, if what it is doing gave rise, historico-evolutionarily,

to a reproductive advantage.18 The producing system gives rise,

via the consuming system, to the timidity reaction, and is thus

doing what it’s supposed to be doing.

The problem for Millikan becomes apparent when we add

another stipulation to the Bog Scenario. The producing system of

18 Macdonald, G. and Papineau, D. “Prospects and Problems for Teleosemantics” in Teleosemantics 11.

Baseley 17

bogs often represents falsely; what is represented as being in

reality is not really in reality. Shadows are represented as

dangers, but on no obvious account of ‘danger’ are shadows

included; shadows per se do not pose direct threats. But this is

fine, on Millikan’s view. Things get murky when we allow the bog

to represent veridically. In terms of executing biological

functions, we begin to see an important difference between the

bogs’ veridical representation of shadows and the bogs’ false

representation of shadows. As has been shown, false

representations of shadows allow the bog to direct behavior

affectively in pursuit of some biological end. But not so for

veridical representations.

Representations of shadows as shadows would, presumably, not

trigger the right kinds of responses from the consuming system of

bogs; shadows as shadows would not alert the bog to any potential

threat. Assume that nine out of ten times, no predator lurks in

the shadows. On the basis of this, one can say that shadows, for

the most part, do not constitute a danger; shadows are neither

dangerous in themselves nor do they reliably signal the presence

of danger. Thus, a false representation of shadows involves

Baseley 18

shadows being dangerous; and a true representation of shadows

involves no element of danger.

The bogs’ consuming system interprets the producing system’s

shadow ‘image’ as a signal for danger. Given the biological and

physiological makeup of the bog, it can afford to be cautious; as

I said, it can go days and days without eating. Thus, the right

kinds of behavior, i.e., fleeing from shadows, would be directed

toward some biological end, if the bog reacts to shadows

cautiously. That is, despite the fact that nine out of ten times

nothing lurks in the shadows, the bog reacts to shadows as if

they were dangerous. Reacting in this way leads to the bogs’

survival and inclusive fitness. This entails no biological

malfunctions. For even though shadows are represented as being

dangerous—in reality shadows are not—bogs suffer no biological

consequences for being overly cautious. This leads one to

question the biological function of producing systems.

What is more, accurate representations of reality would not

contribute to the organism’s success; only false, or illusory,

representations would tend toward the organism’s biological

success. An accurate representation of shadows entails nothing

Baseley 19

very special: shadows are dark areas produced by obstructions of

light, not areas in which predators lurk. Shadows, understood

this way, would not elicit the right kind of response from bogs.

The right kind of response is to flee, and non-dangerous shadows—

or shadows as they really are—would not elicit a fleeing

response. Hence, the function of producing systems is to produce

representations that lead to survival—regardless of whether the

representations are veridical. On an intuitive level, this

conclusion seems quite odd indeed, leading one to question

Millikan’s position.

A Possible Response from the Millikan Camp

Perhaps we’re getting ahead of ourselves. If, as in the Bog

Scenario, bogs consistently represent shadows as threats—if, that

is, shadows are interpreted by the consuming system as threats

and respond to them appropriately—there would have to be a good

reason for this reaction and behavior. Shadows do not seem to—in

and of themselves—pose direct threats. Provided it’s not the

nature of representational systems to represent falsely and

provided there’s no natural link between false representations

and organisms’ successes — two fairly uncontentious assumptions—

Baseley 20

there must be something about shadows specifically that gives the

bog good reason to act towards them in the way bogs do. Can

every sustained—i.e., non-sporadic, patterned—false

representation be explained by the history of an organism’s

relationship with the thing being represented? Perhaps.

Certainly the bog’s representing shadows as dangerous can be

explained by an analysis of the bog’s predecessors’ history vis-

à-vis shadows. For instance, say the bog’s predecessors’ primary

predators were those that resided in, and only attacked under,

the cover of darkness. The bogs that tended to survive were

those that avoided all shadowy, dark things. Caution vis-à-vis

shadowy things was selected for, since its presence lead to the

survival of more animals with that trait. Hence, today there are

bogs that respond cautiously to shadows; they have good

historico-evolutionary reasons for doing so. The creatures

respond to features of their current environment in the same way

that their ancestors would have responded.

The broader question, one alluded to in the previous

section, is whether false representations that support organisms’

biological successes is a problem for Millikan’s teleosemantics.

Baseley 21

Provided that Millikan is willing to decouple veridical

representations and consuming system’s proper behavior, she can

walk away from the Bog Scenario unscathed. Given the Bog

Scenario, we can answer this question negatively; falsidical

representations that support organisms’ successes are not a

problem for Millikan, despite intuitions to the contrary.

Indeed, at this level teleosemantics has no strong loyalty to

veridical representations, provided non-veridical representations

are accounted for historically. If explanatory pride of place is

given to an organism’s history and biological functions, then

history itself can account for an organism’s false

representations, so long as it can be shown that falsidical

representations led to that organism’s success. Thus we conclude

that teleosemantics is not concerned with truth or veridicality,

per se, but with an organism’s history.

Concluding Remarks

I’ve attempted to give a detailed description of Ruth

Millikan’s teleosemantics. Her programme began in the mid-1980s

with her book Language, Thought, and Other Biological Categories. From

then, she made an important and novel distinction between

Baseley 22

producing systems and consuming systems of representation. I

have described these and given examples of them, both inside and

outside of cognitive systems. After laying out the groundwork of

Millikan’s approach, I took a critical position, asking whether

Millikan’s account of false or illusory representations posed a

problem for her teleosemantics. Ultimately, I concluded that it

did not; teleosemantics can get off the ground and, indeed, stand

on quite firm footing, even if false representations are required

by organisms for survival. Millikan’s teleosemantics is not

concerned with truth, per se, but with the history of an

organism’s functions and biological success.

Baseley 23

Works Cited

Adams, Fred. “Thoughts and Their Contents: Naturalized

Semantics.” In Philosophy of Mind

(Ed. Stephen Stitch and Ted Warfield.) Oxford: Blackwell

Publishing. 143.

Garson, Justin. The Biological Mind. New York, New York: Routledge.

137-40.

Lycan, William. “The Mind-Body Problem.” In The Philosophy of Mind

(Ed. Stephen Stitch and

Ted Warfield.) Oxford: Blackwell Publishing. 55.

Baseley 24

Macdonald, G. and Papineau, D. “Prospects and Problems for

Teleosemantics.” In

Teleosemantics (Ed. Graham Macdonald and David Papineau.)

Oxford: Clarendon Press.

3-11.

Millikan, Ruth. Language, Thought, and Other Biological Categories.

Cambridge, MA:

MIT Press

Millikan, Ruth. “Useless Content.” In Teleosemantics. Oxford:

Clarendon Press. 100.

Pietroski, Paul. “Intentional and Teleological Error.” Pacific

Philosophical Quarterly, 73: 267-

82.

Shea, Nicholas. “Millikan’s Contribution to Materialist

Philosophy of Mind.” Matiere Premiere

(2006): 127-156.

Wright, Larry. “Functions.” Philosophical Review 82: 139-168.

Baseley 25