Sei sulla pagina 1di 10

Rock Beats Scissors: Historicalism Fights Back

Author(s): Fred Adams and Ken Aizawa


Source: Analysis, Vol. 57, No. 4 (Oct., 1997), pp. 273-281
Published by: Oxford University Press on behalf of The Analysis Committee
Stable URL: http://www.jstor.org/stable/3328087
Accessed: 04-04-2016 03:01 UTC

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

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

The Analysis Committee, Oxford University Press are collaborating with JSTOR to digitize,
preserve and extend access to Analysis

This content downloaded from 168.176.5.118 on Mon, 04 Apr 2016 03:01:40 UTC
All use subject to http://about.jstor.org/terms
Rock beats scissors: historicalism fights back
FRED ADAMS & KEN AIZAWA

1. Introduction

Jerry Fodor (1994) thinks that content is not historically determined. In


this paper we will consider Fodor's reasons. One of us accepts historical-
ism, the other is neutral, but we both think that Fodor's arguments are
unsuccessful.
A good example of an historical account of content is Dretske's (1988).
On this theory, an item C means F when C has acquired the function of
indicating Fs. What C has the function of indicating will depend upon the
environment in which C acquires its indicator function. C is a part of
system S, and S is in environment El. Were one to have put S in E2, C may
have acquired the function of indicating Gs, not Fs. C would then mean G,
not E This type of account is historical because it makes C's meaning
depend upon S's history (the environment S is in and the indicator function
C acquires in that environment). It is also historical because it makes C's
meaning depend upon actual events. C comes to mean F in virtue of actual
episodes of indicating Fs and in virtue of that indication's causally explain-
ing relevant bodily movements.
Other accounts of content are ahistorical. For example, on Fodor's
(1990) theory of content, an item 'X' means X when it's a law that 'Xs
cause "X"s' and nothing else would cause 'X's, but for the fact that Xs do.1
This account is apparently ahistorical for the reason that laws are ahistor-
ical - or at least, the laws upon which content depends are supposed to be
ahistorical. Suppose that 'X' is a syntactical item in a system S. Whether
Xs would cause 'X's is not supposed to depend upon S's history. Also, for
Fodor, whether 'X' means X is independent of any actual events where Xs
causes 'X's. Counterfactuals suffice (or so it is claimed).

2. More Dretske

On Dretske's account there are at least three ways history matters. In El


there may be no Gs. Thus, C tokens can only indicate Fs (not Gs). C comes
to acquire the function of indicating Fs, when C's indicating Fs causally
explains some relevant bodily movement M of system S (a system in which

1 We are well aware that in 1990 Fodor had two versions of his asymmetrical causal
dependency theory - the mixed view and the pure view. Subsequently he abandoned
the mixed view because of its 'historical' component, viz. in the mixed view the law
that Xs cause 'X's had to be instantiated.

ANALYSIS 57.4, October 1997, pp. 273-281. 0 Fred Adams and Ken Aizawa

This content downloaded from 168.176.5.118 on Mon, 04 Apr 2016 03:01:40 UTC
All use subject to http://about.jstor.org/terms
274 FRED ADAMS & KEN AIZAWA

C is contained). However, were S in a different environment E2, there may


exist both Fs and Gs. C may indicate the presence of both Fs and Gs, when
tokened. However, C may still come to mean only G (not F), when C
acquires the function to indicate Gs (not Fs). C does this when it is C's indi-
cation of Gs (not Fs) that explains C's causing relevant bodily movement
M. Thus, one way the environment (and history) can matter is that the
environment may not contain Gs for C to indicate or mean. And this may
be so even if it is true counterfactually that were there to be Gs in El, C
would indicate them (whether or not C would come to mean G), when
tokened. A second way that history matters is that in E2, C's indicating Gs
may explain why C caused relevant bodily movement M. However, in
another environment E3, C's indicating Gs may not explain its causing of
relevant bodily movement M. What C's indication of Gs explains also may
depend upon the environment S is in. Thus, not only is what C indicates
environment-relative, but also what C's indication explains. Those are two
ways in which history matters on Dretske's account. A third and most
important way that history matters, in contrast to Fodor, is that for Dret-
ske there must be some actual events of C's indicating Fs for C to come to
mean E This is a main sense in which an account of content can be histor-
ical and the main sense that is at issue for Fodor. It is the sense that we will
discuss below.

3. More Fodor

Why does Fodor think that 'Where nomic relations are the issue, actual
history drops out and what counts is only the counterfactuals' (Fodor
1994: 115)? There seem to be four reasons, none of which are compelling.
First, cluttering up the pure non-historical theory with historical condi-
tions results in a 'mixed theory' which is 'unaesthetic' (Fodor 1994: 116).
Now we are all for aesthetics. Thus, we cannot resist pointing out that
Fodor's theory is already unaesthetic because of his treatment of demon-
stratives. To wit: 'I suspect, in fact, that it is only demonstrative thoughts
whose content is determined by their actual etiology' (Fodor 1994: 119).
It is unaesthetic to have an historical theory of demonstratives and an ahis-
torical theory of names and kind terms.
While we're on aesthetics, there is an unsightly way Fodor tries to handle
names. Earlier (Adams and Aizawa 1994: 229) we pointed out that Fodor's
semantic theory would not easily accommodate the meanings of names.
His is a theory of laws. Names name individuals and individuals cannot
feature in laws. To get around this Fodor now says that 'Aristotle' means
Aristotle in his mouth because of a nomic relation between the property of
being Aristotle and his property of being (tenselessly) disposed to cause
'Aristotle' tokens in Fodor (Fodor 1994: 118-19). This should make

This content downloaded from 168.176.5.118 on Mon, 04 Apr 2016 03:01:40 UTC
All use subject to http://about.jstor.org/terms
ROCK BEATS SCISSORS: HISTORICALISM FIGHTS BACK 275

'Aristotle' mean the property of being Aristotle, not Aristotle. (Bear in


mind, Aristotle is dead, while the property of being Aristotle is neither
dead nor alive ... different categories ... different things to be meant.)
Also, don't overlook Fodor's appeal to the role of experts in mediating
his use of 'Aristotle' and the property of being Aristotle. Fodor seems to
think there is a role for the experts to play here as there is a role for the
experts to play in connecting his use of 'elm' with the property of being an
elm. But what role can there be for the experts, if there exists a nomic
relation between 'Aristotle' and the property of being Aristotle, whether or
not that law is instantiated (in Fodor, or in anyone)? Now Fodor suggests
that the experts mediate the connection between 'Aristotle' and the prop-
erty of being Aristotle. But appealing to experts smacks of historicalism in
a number of ways. Here's why. Experts are historical beings. Interacting
with experts is an historical event. Also, what one means by a term will
depend upon the experts with which one interacts (an historical event). For
suppose Gary overhears a conversation that involves use of a term
'Aristotle'. On the basis of this Gary acquires the term. Now the experts
are the persons who can discriminate Aristotle. But there could be more
than one conversation. Gary could have overheard philosophers discussing
Aristotle's metaphysics. Or he could have overheard drunks talking about
a dog named 'Aristotle' living in New Jersey. What Gary means when he
tokens 'Aristotle' (the philosopher or the dog) will depend upon the
experts ... that is, which group Gary actually overheard.2 This certainly
looks historical in a number of ways.
A second reason Fodor offers against historicalism is a putative auto-
nomy of content (Fodor 1994: 89-90). Fodor tries to employ a familiar
Chomskian line about the stimulus independence of language and thought.
As Chomsky (1959) correctly observed, presented with a stimulus of a
painting, one might respond in an open-ended number of ways. 'It's hung
too high.' 'Clashes with the carpet.' 'Dutch.' Similarly, presented with a
chicken, there are a number of things one can think. 'Why did this thing
cross the road?' 'Future ingredient for noodle soup.' 'Dinner!' Chomsky's
point is that what you think or say now is not determined by your environ-
mental stimulus now. This seems fine and good. Nevertheless, it does
nothing to impugn historicalism. Historicalists can agree that what
happens inside the head depends not only on what is currently impinging
on the senses, but Fodor goes further and tries to draw a consequence from
this that does not follow. From the fact that one may token 'cat' in the
absence of cats Fodor concludes that no token of 'cat' need have actually
been caused by cats for 'cat' to mean cat. This further claim simply does

2 Let it be equally probable that Gary would overhear one conversation or the other.

This content downloaded from 168.176.5.118 on Mon, 04 Apr 2016 03:01:40 UTC
All use subject to http://about.jstor.org/terms
276 FRED ADAMS & KEN AIZAWA

not follow. A historical theory may maintain that current meaningful


tokens of 'cat's need not actually be caused by cats, even though in the past,
some tokens of 'cat's must have been caused by cats for 'cat' to acquire its
meaning (but not all tokens and not for ever). No historical theory we
know maintains that for 'cat' to mean cat now it must be caused by a cat
now. 3
Perhaps Fodor is suggesting that since behaviour is stimulus independ-
ent, so is content. For he says: 'It's a difference between cats and rocks that
the trajectories of the former are not, in general, predictable from the
currently impinging environmental forces' (Fodor 1994: 89). But still it
does not follow from Chomsky's point, nor from the difference between
cats and rocks (and the predictability or non-predictability of their behav-
iour by currently impinging forces) that content is ahistorical. Behaviour is
stimulus independent when there are other intentional states that cause
behaviour independently of (or in concert with) stimulus conditions.4
Content is stimulus independent once a concept is acquired.5 Once you've
got the concept of a cat, 'cat' tokens can be caused in a myriad of ways.
None of this shows that the acquisition of a concept is independent of
causal history of all tokens.
Fodor sets out a third reason against historicalism in the following
passages:

Your behavior typically outruns your experience. This is because the


causation of much of your behavior is mediated; much of your behav-
ior is caused by your thoughts and the content of your thoughts typi-
cally outruns your experience. The content of your thoughts can
outrun your experience for the same reason that your thoughts can be
stimulus free: although content is a causal notion, it's possible rather
than actual etiology that counts semantically. Roughly, according to
informational semantics, to think that it is raining pigs is to have a
thought that you would be caused to have if it were to rain pigs. As
long as the counterfactual is in place, you can think about a rain of
pigs without ever having been in one (Fodor 1994: 90-91).
In this passage, Fodor gives one explanation of why the contents of
your thoughts typically can outrun your experience. Another story we
have heard (from Fodor, no less) involves the language of thought. The
reason you can think of a rain of pigs without ever having been in one is
that you have the concept of rain, the concept of pigs, and you can

3 This is certainly not a feature of Dretske's historical account, for example.


4 This is why functionalism was superior to behaviourism.
s That is why concepts can be falsely tokened.

This content downloaded from 168.176.5.118 on Mon, 04 Apr 2016 03:01:40 UTC
All use subject to http://about.jstor.org/terms
ROCK BEATS SCISSORS: HISTORICALISM FIGHTS BACK 277

concatenate concepts. This is true even if to have the concept of rain one
must have experienced rain and to have the concept of pigs one must have
experienced pigs. Productivity alone doesn't impugn historicalism of
content.

A fourth reason Fodor offers against historicalism is that evidence for


historical determinants is 'equivocal'. Here Fodor maintains that intuitions
over Davidson's Swampman are stronger than those over Twin-Earthian
considerations favouring an historical account of content. This is probably
more autobiographical of Fodor's intuitions than an argument of any
weight, but here goes.

4. Swampman
Davidson (1987) gives the example of Swampman as follows. Davidson
is vaporized by lightning. At exactly the same time and out of new mole-
cules Swampman is miraculously created and is Davidson's physical
replica. Swampman has all of the same physical movements that David-
son would have, seeming to recognize Davidson's friends, seeming to speak
English. No one can tell Swampan from Davidson. Davidson maintains
that Swampan would have no thoughts or intentional states at all,
nonetheless.

My replica can't recognize my friends, it can't re-cognize anything


since it never cognized anything in the first place. It can't know my
friends' names (though of course it seems to), it can't remember my
house ... Indeed, I don't see how my replica can be said to mean
anything by the sounds it makes, nor to have any thoughts (Davidson
1987: 433-44).
Of the Swampman example, and Davidson's historicalism, Fodor says
this:

One might, like Davidson, treat these consequences of the historical


account of content as bullets that one's intuitions must learn to bite.
But I think they should be viewed rather as serious embarrassments
for Davidson's causal history kind of externalist semantics. Of course,
not having had one, Swampman doesn't remember his twelfth birth-
day party; 'remember' is factive, and you can't remember what didn't
happen. But it seems very odd to say that Swampan doesn't know
what time or day of the week it is, since (to put it in a way that of
course begs the question) if you ask him what time of day it is, he is
perfectly able to tell you .... I think the unbiased intuition is that
Swampman thinks all sorts of things that Davidson does: that it's
Wednesday, and that radical interpretation is possible, and that water

This content downloaded from 168.176.5.118 on Mon, 04 Apr 2016 03:01:40 UTC
All use subject to http://about.jstor.org/terms
2.78 FRED ADAMS & KEN AIZAWA

is wet, for example. I think this because, although he lacks Davidson's


causal history, Swampman shares Davidson's dispositions, and it's the
counterfactuals that count for content, just as informational theories
claim (Fodor 1994: 117-18).
On the surface it appears that Fodor thinks that he has intuition on his
side. Fodor believes that intuition says that Swampman has thoughts, and
if historical accounts do not, then so much the worse for historical
accounts.

What Fodor finally says is somewhat weaker. He says that 'It's intuitively
plausible that he [Swampman] has states that are their [thoughts]exact
ahistorical counterparts and that these states are intentional' (Fodor 1994:
117). This is weaker because what could be running through Swampy's
head are syntactic items that are semantically uninterpreted. These may be
counterparts of actual thoughts. The dispute is over whether or not they
indeed are intentional.
At one point Fodor asks 'If it's not [Swampy's] believing that it's
Wednesday that explains why the Swampman says "It's Wednesday" when
you ask him, what on earth does'? The answer is that his having the syntac-
tic items 'It is Wednesday' in his belief box explains it. However, this would
explain it even if what is in Swampy's belief box were semantically
uninterpreted. Compare a similar question posed to Feigenbaum about
Eliza ... 'If Eliza doesn't really believe that it is Wednesday, then why does
she say "It is Wednesday", when asked what day it is?' Surely the answer
Fodor would give for Eliza would not be that Eliza believes it is Wednes-
day. Similarly, we see no good reason why our intuitions must say that it is
because Swampy believes that it is Wednesday that the words 'It is Wednes-
day', come out of his mouth, when presented with 'What day is it?'
Swampy may be physically and syntactically driven just as surely as is
Eliza.
Of course, Eliza is not physically identical to a system with genuine
thoughts. So this cannot be the whole story. But it can be part of the answer
- the part that relies on syntax in place of semantics to explain bodily
movements that are identical to movements that are part of intentional
behaviour (Adams and Fuller 1992, Adams, Fuller, and Stecker 1993).
Swampy's bodily movements in uttering 'It is Wednesday' need not be
construed as intentional behaviour (being not caused by intentional states),
even though they are identical to Donald's bodily movements in his inten-
tionally saying 'It is Wednesday' (being caused by intentional states)
(Dretske 1988, Adams, et.al. 1990, Adams 1991).
At another point Fodor explains his intuition that Swampy means water
by 'water' this way:

This content downloaded from 168.176.5.118 on Mon, 04 Apr 2016 03:01:40 UTC
All use subject to http://about.jstor.org/terms
ROCK BEATS SCISSORS: HISTORICALISM FIGHTS BACK 279

The reason he does [mean water by 'water'] is that it's water that
would cause his 'water' tokens in the worlds that are closest to the one
that Swampman actually lives in. Roughly, a thought means what
would cause it to be tokened in nearby possible worlds (Fodor 1994:
118).
However, the closest possible world to the actual is this world and in this
world 'What day is it?' causes Swampy to token 'Wednesday'. So by this
reasoning, for Swampy, 'Wednesday' should mean 'What day is it?' This
seems to be the wrong result for the meaning of Swampy's 'Wednesday'
tokens (Adams and Aizawa: 1994). So why be compelled by Fodor's intu-
itions about Swampy's 'water' tokens? The question 'What have you got
there in the glass [of water] in your hand?' will cause Swampy to utter
'Water', since it would cause that in Donald (his physically identical twin).
So do Swampy's 'water' tokens mean 'What have you got there in the glass
in your hand?' Surely not!
Of course, Fodor would insist that questions about what is in the glass
would only cause 'water' tokens in Swampy because water would cause
'water' tokens in Swampy. But that is a large part of what is at issue. Intu-
ition cannot settle whether that is true. Indeed, our intuitions are that it is
not true. Construed purely syntactically, there ought to be significantly
many kinds of things other than water (bumps on the head, electrodes in
the brain, etc.) that are physically capable of triggering Swampy's 'water'
tokens, independently of water's triggering them. Indeed, our view is that
Fodor's asymmetric dependencies cannot be the source of such meaning
because, construed purely syntactically, they seem quite implausible
(Adams and Aizawa: 1994).
Fodor also constructs a hybrid twin/swampman example supposedly
showing that intuition is on his side. Here the idea is that there are two
swampmen - Earth Swampy and Twin-Earth Swampy. What do Twin
Swampy's 'water' tokens mean? According to Fodor, they mean XYZ.
They mean this not because of any actual causal connection or causal
history with XYZ, because (Fodor decrees) there is none. They mean this
because in the nearest possible world where Twin-Swampy's 'water' tokens
are tokened XYZ would cause them to be tokened.
First, we disagree that Twin-Swampy's 'water' tokens mean anything (at
least, initially). So we do not share Fodor's intuition. If one thinks that the
original Swampman has no thoughts, there is nothing in the Twin Swamp-
man case that should lead one to abandon the view that content is
historical. Over time, of course, Twin-Swampy is likely to acquire the
concept of twin-water, just as Swampy is likely to acquire the concept of
water. Second, suppose that you think you share Fodor's intuitions to the
extent that if Twin-Swampy's 'water' tokens had meaning they would

This content downloaded from 168.176.5.118 on Mon, 04 Apr 2016 03:01:40 UTC
All use subject to http://about.jstor.org/terms
2.80 FRED ADAMS & KEN AIZAWA

mean XYZ, not H20. What explains that? The counterfactuals alone? Not
necessarily. Contrary to Fodor, Twin-Swampy would mean XYZ if
anything by 'water' because even though in the actual Twin-world he has
no thoughts, in the nearest possible world where Twin-Swampy had
thoughts he would have the same causal histories as his physical twins on
Twin-Earth. Sharing their causal histories would suffice for sharing
'water'-tokens with the same content. Therefore, even if we grant Fodor
that Twin-Swampy's thoughts would mean XYZ, not H20 we need not
grant his view of why they would mean that.

5. Back to Twin-Earth

At this point we cannot refrain from pointing out that, minus an historical
condition, Fodor's 'pure' theory of content cannot solve the original Twin-
Earth problem of content. His historical 'mixed' theory could solve it
because whether Jerry meant H20 by 'water' or Twin-Jerry meant XYZ
was dependent upon which laws Jerry or Twin-Jerry instantiated (an
historical condition). Taking away this historical condition now leaves it
indeterminate that Jerry means H20 by 'water' and not XYZ (or that
Twin-Jerry means XYZ, not H20). After all, Fodor admits of such cases
'... I don't claim that they are impossible, or even that they don't happen
(cf. the familiar story about jade and jadeite)' (Fodor 1994: 30). What
Fodor now says about Twin cases is that they don't destroy the possibility
of a broad-content psychology, as he formerly believed they would. He
now says that Twin cases, instead, would create a failure to express gener-
alizations that subsume such Twins, but that as long as such failures are
few and far between they can be regarded as 'accidents' and as 'spurious'.
That may be a possible way to treat Twin cases with respect to generaliza-
tions and explanations of behaviour, once content laws are in place. But it
is not a way to handle Twin cases in determining the contents of their
thoughts originally. As far as we can tell, Fodor's new theory cannot handle
content assignments for the original Twin cases at all. For without looking
to causal history, there is no more reason to assign H20 than XYZ as the
meaning of Jerry's (or Twin-Jerry's) 'water' tokens.

6. Conclusion

Fodor offers four reasons for thinking that content is not historical. As we
have shown, none of these reasons work. One who thinks content is histor-
ically determined should not be moved by Fodor's arguments. On intuition
alone, people are divided about whether Swampman has thoughts. Histor-
ical theorists (Davidson, Dretske, Millikan et al.) say no. Fodor says yes.

This content downloaded from 168.176.5.118 on Mon, 04 Apr 2016 03:01:40 UTC
All use subject to http://about.jstor.org/terms
ROCK BEATS SCISSORS: HISTORICALISM FIGHTS BACK 2.81

However, Fodor gives us nothing to take us beyond these intuitions. So our


reply to Fodor's attempted rejection of historicalism ('Scissors cuts paper,
Swampman eats Twins' (Fodor 1994: 117)) is this: rock beats scissors and
Twins bite back! 6

University of Delware
Newark, DE 19716, USA
fa@udel.edu

Centenary College
Shreveport, LA 71134, USA
kaizawa@beta.centenary.edu

References
Adams, F. 1991. Causal Contents. In Dretske and His Critics, ed. B. McLaughlin, 131-
56. Oxford: Basil Blackwell.
Adams, E, D. Drebushenko, G. Fuller and R. Stecker. 1990. Narrow content: Fodor's
folly. Mind and Language 5: 213-29.
Adams, F. and K. Aizawa. 1994. Fodorian semantics. In Mental Representation, ed. S.
Stich and T. Warfield, 223-42. Oxford: Basil Blackwell.
Adams, E and G. Fuller. 1992: Names, contents, and causes. Mind and Language 7:
205-21.

Adams, E., G. Fuller and R. Stecker. 1993. Thoughts without objects. Mind and
Language 8: 90-104.
Chomsky, N. 1959. Review of Skinner's Verbal Behavior. Language and Philosophy 35:
26-58.

Davidson, D. 1987. Knowing one's own mind. Proceedings and Addresses of the Ameri-
can Philosophical Association 60: 441-58.
Dretske, E 1988. Explaining Behavior. Cambridge, Mass.: MIT Press.
Fodor, J. 1990. A Theory of Content and Other Essays. Cambridge, MA: MIT Press.
Fodor, J. 1994. The Elm and the Expert. Cambridge, Mass.: MIT Press.
Putnam, H. 1975. The Meaning of 'Meaning'. In Language, Mind and Knowledge, ed.
K. Gunderson, Minnesota Studies in the Philosophy of Science 7. Minneapolis:
University of Minnesota Press.

6 We thank the editor of this journal for careful reading and useful suggestions.

This content downloaded from 168.176.5.118 on Mon, 04 Apr 2016 03:01:40 UTC
All use subject to http://about.jstor.org/terms

Potrebbero piacerti anche