Kanzi Revisited

First, consider the bonobo Kanzi's planning, thinking, and decision making that utilizes a 256-symbol keyboard, which Kanzi has, over the years, learned to use to communicate his beliefs and desires. Bonobos in the wild surely have desires (e.g., for bananas), but just as surely they don't have the kind of sophisticated, cooperative desires that Kanzi expresses, such as the desire to be taken by a particular person to a sequence of locations, or to do one activity first and then another. What the symbol board has done is to reconfigure Kanzi's capacity for belief and desire, much as our using pen and paper reconfigures our mathematical ability by augmenting the in-the-head capacity we have for multiplication. Both are cases in which an external symbol system becomes integrated with preexisting cognitive capacities in ways that significantly modify the nature of those capacities. We can, of course, distinguish between the parts of those capacities that are internal and those that are not, but this is already to concede that the overall cognitive process itself is extended. (Wilson, 2010, p. 180).

I'm going to guess that lodging a bullet in the brain would reconfigure Kanzi's capacity for belief and desire.  And it would be pretty well integrated with preexisting cognitive capacities in ways that significantly modify the nature of those capacities.  But, I wouldn't want to say that Kanzi's cognitive processes extend into the bullet.  (Maybe it would be a cognitive resource, but regarding that idea see my post on "Cognitive Resources".) 

This looks to me to be just a thinly veiled version of the coupling-constitution fallacy.  To reconfigure is to have a certain kind of causal influence.  To be integrated is to have a certain kind of persistent causal influence.

Maybe Wilson could add, following Clark, that the reconfiguring, integrated thing has to be an information processing resource.  How about embedding a measuring tape in Kanzi's brain?  (Maybe that would be a cognitive resource, but ... see above.)

Well, maybe, if the measuring tape were integrated in the right way in the overall information processing economy of Kanzi's brain, then cognitive processing would extend into it.  If, by this, Wilson means that, were the measuring tape to bear non-derived content and be manipulated in particular sorts of ways characteristic of cognitive processing, then yes.  That's the Adams and Aizawa view.  Our point is that such reconfigurations and integrations do not typically lead to the right sorts of information processing economies.   But, in any event, this line does not seem to be the sort of reconfiguration and integration Wilson has in mind, since he doesn't mention them.

Do we have a winner?

So, it is back to Wilson's "Meaning making ..." where I think we find an instance of the C-C fallacy.

Common to both active cognition and cyborg fantasy arguments for extended cognition is the idea that by examining just what is involved in the exercise of some particular cognitive capacity, one finds that it actually does or could well involve causal loops that extend beyond the body of the individual agent. In particular, these causal loops (do or may) pass through objects and other entities in the agent's environment, and it is only the whole, functioning, beyond-the-head causal system that constitutes the matter in motion that realizes the exercise of the capacity. (Wilson, 2010, p. 173).

So, first of all, it's nice to have another example of the advocates of EC invoking the language of causation and constitution.  Further support for my contention that Ross and Ladyman should be directing their critique not just at Adams and Aizawa, but at least at a large number of advocates of EC.

Second, in the second sentence, it looks to me as though Wilson is moving from an observation about the causal contribution of some extracortical processes to cognitive processing to a conclusion about the constitutive contribution of these extracortical processes. What else could be going on?

... Adams and Aizawa (2001) on Clark and Chalmers (1998), Grush (2003) on Haugeland (1998), and, most recently, Rupert (2004) largely on Rowlands (1999). For the most part, these critiques have to reconstruct, sometimes quite imaginatively, the arguments that they critique, leaving one with the feeling that externalists must surely have something more up their sleeves than what their critics draw from the hat. (Wilson, 2010, p. 173)

And I agree that one has to use some imagination to try to figure out what the EC arguments are supposed to be.  But, the only thing that seems to me to make sense of the text is the C-C fallacy.  Of course, that's not a good argument, which is a defeasible reason to think that the reading is incorrect.

But, then again, "philosopher makes mistake" is not news, right?  Yes, it would be great to think that the EC folks have something more up their sleeves, but what is it?  Adams and I have also taken into consideration the stuff about loops, "trust and glue", etc., etc.  I feel as though we are still waiting.

Ross & Ladyman's "Alleged Coupling-Constitution Fallacy" 2

extended mind thesis and its internalist rival can be interpreted as alternative claims about what kind of ontology of systems any cognitive model should presuppose.  The idea that there might be a justified general such claim about all cognitive models, which could rationally be made in advance of tackling specific modeling problems one at a time, would have to be based either on a universal tractability constraint or on metaphysics. (Ross & Ladyman, 2010, p. 156).

Two things here.

I) I don't see the internalist/externalist debate as, in the first instance, offering alternative claims about what kind of ontology of systems any cognitive model should presuppose.  There is some question about what the ontology of cognition is--e.g., whether cognition involves representations--but once that is settled the core question becomes where in the world one finds those bits of ontology.  So, ontology is, at most, only part of what is in play in the EC debates.

II) I think that Ross and Ladyman are missing what is going on in the EC literature.  They refer to this "idea that there might be a justified general such claim about all cognitive models, which could rationally be made in advance of tackling specific modeling problems one at a time".  But, Rupert, Adams and I take it that there has been this enterprise of cognitive psychology that has been going on for some decades and that one can extract from that some basic conclusions about cognitive ontology.  In fact, I think that most parties to the EC debate probably agree to the method of abstracting conclusions about cognitive ontology from actual scientific practice. (Maybe not the same scientific practice, mind you.  Some folks extrapolate from Gibson, where others extrapolate from other sources.)  Where parties to the EC debate differ is in what they think is correct to abstract from that actual scientific practice.

Ross & Ladyman's "Alleged Coupling-Constitution Fallacy" 1

This paper began life as an account of how Adams and Aizawa were stuck in the 17th Century with immature (pre-?)scientific metaphysical assumptions about causation and constitution.

After a very enjoyable dinner at King's College, Cambridge, (hosted by Mark Sprevak as part of a workshop on Computation in Cognitive Science), I was able to encourage James to revisit the draft of the paper that I had seen.  I proposed that the problems they had with causation and constitution were not the doing of Adams and Aizawa, but were part and parcel of much of the EC debate.  So, now the paper is about how (almost?) everyone in the EC debate is stuck in the 17th Century with immature (pre-?)scientific metaphysical assumptions about causation and constitution.

Ross and Ladyman should be happy about this revision, since they are now going after more and bigger fish than just Adams and Aizawa.  Adams and I are happy, since now the mess is not our fault.

One of the challenges of responding to the Ross-Ladyman critique, however, is that they are not so much interested in adjudicating the particular issues that arise in the EC debate as they are making the case that everything must go.

Nevertheless, I think that even though they are now dumping on the whole EC debate, they are coming down a bit too hard.  I won't try to "refute" the criticisms so much as try to make it less obvious that the whole EC debate is mired in the useless and primitive.  (A modest goal to be sure...)

Menary(?) and Wheeler(?) Buy the Distinction between Causation and Constitution

Wheeler (chapter 11) argues that the extended mind is a kind of extended functionalism. Wheeler points out that the extended mind is not simply a weak claim about the causal dependence of some cognition on external factors (cf. Adams and Aizawa's coupling-constitution fallacy). It is a stronger claim involving the constitution of cognition, at least in part, by external factors. Therefore, the extended mind is not simply an embodied-embedded thesis that treats external props and tools as causally relevant features of the environment. It is a thesis that takes the bodily manipulation of external vehicles as constitutive of cognitive processes. (Menary, 2010, p. 21).

So, this pretty clearly buys into the causation/constitution distinction in order to make out the hypothesis of extended cognition.  So, as I noted earlier, pace Hurley, it seems to me that it is the job of the advocates of EC, not the critics, to explicate this distinction.

The Ambiguity of "Constitution"

Hurley (chapter 6) and Ross and Ladyman (chapter 7) are concerned about the very nature of the alleged fallacy. Hurley complains that philosophers employ the causal-constitutive distinction, on which the causal coupling fallacy trades, without motivating or explaining the distinction in detail. Ross and Ladyman argue that the distinction itself is not used in mature sciences such as economics and physics. (Menary, 2010, p. 13).

Another important thing to note is that "constitute" has at least two meanings.  "X constitutes a cognitive process" could mean something like X is a part, among perhaps others, that gives rise to a cognitive process.  On this reading, "X constitutes a cognitive process" would be closely related to "X realizes a cognitive process".  Alternatively, "X constitutes a cognitive process" could mean simply that X is a cognitive process.  Compare: That constitutes a good answer = That is a good answer.

So, on this second reading, the coupling-constitution distinction comes down to a distinction between X causing, or being caused by, a cognitive process, versus X being a cognitive process.  The nice thing about the word "constitution" is that it does have this ambiguity, which picks up on the ambiguity I think is to be found in the EC literature.

Here's what I found at www.dictionary.com:

1. to compose; form: mortar constituted of lime and sand.

2. to appoint to an office or function; make or create: He was constituted treasurer.

3. to establish (laws, an institution, etc.).

4. to give legal form to (an assembly, court, etc.).

5. to create or be tantamount to: Imports constitute a challenge to local goods.

I'm thinking #1 and #5 get at the ambiguity I'm describing.

Burden Shifting

Hurley (chapter 6) and Ross and Ladyman (chapter 7) are concerned about the very nature of the alleged fallacy. Hurley complains that philosophers employ the causal-constitutive distinction, on which the causal coupling fallacy trades, without motivating or explaining the distinction in detail. Ross and Ladyman argue that the distinction itself is not used in mature sciences such as economics and physics. (Menary, 2010, p. 13).

I don't take it to be my job to explicate the coupling-constitution distinction that the EC folks use to articulate their view.  That's their job.  They are the ones who say that the environment does not merely causally influence cognitive processes, but that they constitute those cognitive processes.

But, matters are not as hopeless as Hurley might suggest.  There is some discussion of causation versus realization and constitution in, for example, Aizawa, K., and Gillett, C., The (Multiple) Realization of Psychological and other Properties in the Sciences. (2009). Mind & Language, 24, 181-208.  Maybe giving this reply is a mug's game though.  No matter how much we might say in explication of the distinction, it's easy enough for a philosopher to complain that we have not said enough or that what we say is not convincing.

So, really, if the distinction collapses, it's mostly ok by me.  Maybe psychologists can press on without the burdens of rethinking their discipline based on a confused philosophical distinction.

Reciprocal Coupling Again

Although we can identify the relevant components, and factorize them into internal and external components, the nature of reciprocal coupling makes it difficult to study the components as separate systems because they are continuously influencing and responding to one another. They are coordinating with one another to produce behavior.  (Menary, 2010, p. 4).

Now, why would reciprocal coupling per se make it difficulty to study the components (brain and body) as separate systems?  Even by the EC line, isn't this how things have been done up to this point?  There has to be more to the story.  The thermostat and the furnace continuously influence and respond to one another, but they look to be separable systems (or at least separable sub-systems or components of a single system).

Menary on Symmetric Versus Asymmetric Causal Coupling

In his introduction to The Extended Mind, Menary offers this explication of Clark and Chalmers' causal coupling.

There are two possible interpretations of causal coupling here, and it is important to be clear about which one is implied by C&C.

A. Asymmetric influence: environmental features have a causal influence over inner processes. It may still be the case that we can change the exter­nal environment and that affects competence and behavior of the subject. If you take my diary away from me I won't be able to remember all my engagements. The diary prompts my recall of memories, but there is no need to go further and say that because the diary has a causal influence on me that it is thereby part of my memory, or the cognitive processes that allow me to remember. It would be a mistake to make this claim simply on the basis of a causal connection.

B. Symmetric influence: the inner and outer features have a mutually constraining causal influence on one another that unfolds over time. It is not simply that the diary prompts or causes, as input, various cognitive processes to unfold in my brain; rather, the external process of retrieving the information from the diary and the concurrent processes in my brain jointly govern my future behavior. This is what Menary calls cognitive integration (2006, 2007, this volume).  

This distinction is important because, critics of EM, such as Adams and Aizawa (2001, this volume) are working with interpretation A, but Clark is working with B. Hence there is a misunderstanding between exponents of EM and their critics. The critics may wish to claim that although EM is supposed to endorse B, very often A is what is endorsed, and A is not a constitutive thesis. Exponents of EM must be careful to indicate when and why they are using interpretation B. (Menary, 2010, pp. 3-4)

Ok.  I don't see that reciprocal coupling matters at all.  Take the original story about the bimetallic strip in a thermostat.  You get reciprocal coupling there, but not extended strip expansion.

To begin, we may observe that the mere causal coupling of some process with a broader environment does not, in general, thereby, extend that process into the broader environment. Consider the expansion of a bimetallic strip in a thermostat. This process is causally linked to a heater or air conditioner that regulates the temperature of the room the thermostat is in. Expansion does not, thereby, become a process that extends to the whole of the system. It is still restricted to the bimetallic strip in the thermostat.  (Adams and Aizawa, 2001, p. 56)

Another Instance of the C-C Fallacy?

Here Chemero and Silberstein seem to me to attribute the C-C fallacy to Clark:

Clark (2003) takes this further, arguing that external tools (including phones, computers, language, etc.) are so crucial to human cognition that we are literally cyborgs, partly constituted by technologies (Chemero & Silberstein, 2008, p. 5).

The fallacy here is in thinking that something that contributes "crucially" to some sort of cognitive or behavioral success is thereby cognitive.  The crux of the difference between what Rupert calls HEC and HEMC is the difference between crucial causal contributors being cognitive crucial causal contributors and non-cognitive crucial causal contributors.

Di Paolo on Extended Mind 1

Di Paolo begins his paper,

In a recent article, Wheeler (in press) has put a question mark on the relation between enactivism (in its life/mind continuity version, e.g., Varela et al. 1991; Thompson 2007) and the extended mind (EM) hypothesis (Clark and Chalmers 1998). His conclusion spells gloom for the prospects of a unified non-Cartesian cognitive science: enactivism and EM are demonstrably incompatible!

This conclusion seems at odds with the spontaneous understanding of enactivism as proposing a view of cognition as fundamentally embodied and situated and the (apparently!) parallel understanding of EM as signalling how much of our cognitive skills rely crucially on the availability of non-biological epistemic technologies.  (Di Paolo, 2009, p. 9. italics added).

Now, as I understand it, EM says more than that our cognitive skills rely crucially on the availability of non-biological epistemic technologies.  This sounds like the causal claim that cognition is causally influenced by non-biological epistemic technologies.  But, EM makes the stronger claim that cognition is constituted, in part, by the available non-biological epistemic technologies.  Or, to put the matter in another way, the phrase "rely crucially on" is at best ambiguous between a causal and a constitutive reading.

Avoiding these ambiguities is one reason to keep a distinction between causal and constitutive claims front and center in the EC debates.

Why is the C-C Fallacy So Common? 3

In a recent post on Rowlands, 2009, I noted that it is common to find in the EC literature the view that when some thing is coupled "in the right way" to the mind, then the thing becomes a part of one's cognitive system.  This, however, sometimes leads to the C-C fallacy.  But, what encourages perseverance in the fallacy is, in part, how one understands "in the right way".  For the cognitivist, "in the right way", means something like in the right causal economy of a computation.  But, for Clark, for example, "in the right way" means something about "trust and glue".  So, one might conjecture that part of what enables Clark to commit the C-C fallacy is not appreciating the difference between the two ways of cashing out "in the right way".

Again, this is more speculation on the source of the C-C fallacy.

Why is the C-C Fallacy So Common? 2

Again, assuming the fallacy is common, why is it so common?

Think about the human auditory system (as described in, say, Wikipedia).  It looks like the auditory system is just the set of things that enable a person to hear, e.g., the pinnas of the ears, the auditory canal, the bones of the middle ear, the cochlea, the hair cells, the auditory nerve, etc.  If you take this picture seriously, then it is a small step to saying that one's hand cupped behind one's ear is part of one's auditory system.

If that is one's picture, then maybe the C-C fallacy isn't just a simple fallacy.  Maybe it is that, when pressed on further examples, your view of systems leads to incorrect conclusions.

Why is the C-C Fallacy So Common? 1

This question, of course, presupposes that the fallacy is in fact common, but some support for this is given in The Bounds of Cognition, Chapter 6, and in earlier posts here on the blog.  So, let's go with that.  But, if the fallacy so common, and as some have noted, obviously fallacious, then why do so many philosophers commit the fallacy?  I think there may be many reasons, but here is one.

Compare the burden of proof involved in the simple coupling arguments, on the one hand, versus the burden of proof in cognitive equivalence arguments, on the other.  It is much easier to show that something in the body or environment causally influences what goes on in the mind than it is to show that some transcranial or transcorporeal process is the same as some intracranial cognitive process. There are very few instances in the extended cognition literature where one finds side-by-side comparisons of putative cognitive equivalence like those found in the three modes of Tetris play and in the Inga-Otto case.  Maybe these are the only such comparisons.  Moreover, these sorts of comparisons almost invariably invite observations of dissimilarities.   Such observations are standard among critics of extended cognition, but there appear to be no cases in which a critic denies that the claim that one or another thing in the body or environment causally influences the mind.  So, one can see that advocates of extended cognition will regularly be tempted to claim that causation is sufficient for extended cognition, since it is easy to find causal influences on cognition.  If causation were sufficient for extended cognition, the case for extended cognition could more easily be made

Looks like the C-C Fallacy to Me

Noë seems to me to commit the coupling-constitution fallacy in the move from the second sentence to the third:

I agree with the philosophers Andy Clark and David Chalmers that there is no principled reason not to think of the wristwatch, the landmarks, the pen and paper, the linguistic community, as belonging to my mind.  The causal processes that enable us to talk and think and find our ways around are not confined to what is going on in our skulls.  But that is just a way of saying that the machinery of the mind itself is not confined to the skull. (Noe, 2009, p. 82)

What C-C Fallacy 8

The external features in a coupled system play an ineliminable role—if we retain internal structure but change the external features, behavior may change completely.  The external features here are just as causally relevant as typical internal features of the brain.  (Clark & Chalmers, 1998, p. 9)

Again, the emphasis is on causal connections.

What C-C Fallacy? 7

In the cases we describe, by contrast, the relevant external features are active, playing a crucial role in the here-and-now.  Because they are coupled with the human organism, they have a direct impact on the organism and on its behavior.  In these cases, the relevant parts of the world are in the loop, not dangling at the other end of a long causal chain.  Concentrating on this sort of coupling leads us to an active externalism (Clark & Chalmers, 1998, p. 9)

Again, it looks like the existence of a coupling relation is what Clark and Chalmers think justifies extended cognition.

What C-C Fallacy? 6

Then a bit later in "The Extended Mind," Clark and Chalmers write,

In these cases [of different modes of Tetris play, use of pencil and paper, etc.], the human organism is linked with an external entity in a two-way interaction, creating a coupled system that can be seen as a cognitive system in its own right.  All the components in the system play an active causal role, and they jointly govern behavior in the same sort of way that cognition usually does.  (Clark & Chalmers, 1998, p. 8)

The first sentence indicates that a two-way (causal?) interaction between a person’s mind and certain environmental items, such as a video screen, pencil, and paper and so forth creates a coupled system.  Clark and Chalmers, however, apparently take the existence of extended cognitive systems to be sufficient to establish that cognitive processes extend.   The second sentence backs off of this a bit, noting both that all the components in the system play an active causal role and that there is some kind of cognitive equivalence.

What C-C Fallacy? 5

In running through their three sorts of Tetris play, Clark and Chalmers invoke considerations of something like cognitive equivalence, but they begin the paragraph after that by turning to causal influences:

The kind of case just described is by no means as exotic as it may at first appear.  It is not just the presence of advanced external computing resources which raises the issue, but rather the general tendency of human reasoners to lean heavily on environmental supports. (Clark & Chalmers, 1998, p. 8)

In trying to show that the Tetris case is not exotic, they appeal to humans “leaning” on environmental supports.  This, however, looks like a metaphor for the causal contribution of environmental factors.

What C-C Fallacy? 4

Bounds (pp. 89-91) has a number of quotes which at least suggest that there is a relatively short path from observations of causal environmental influences on cognitive processes to extended cognition processes, but how about reviewing a longer bit of text by Clark and Chalmers?  In the early sections of their paper their appear to be two lines of reasoning meant to support EC.  One is based on causal contribution; the other is based on some sort of "cognitive equivalence" between intracranial processes and transcranial processes.

So, the last sentence of section 1 is "We advocate a very different sort of externalism: an active externalism, based on the active role of the environment in driving cognitive processes" (Clark & Chalmers, 1998, p. 7).  One plausible interpretation here is that there is a coupling-constitution argument being given here.