8/24/07

Ismael, Jenann - Saving the Baby: Dennett on Autobiography, Agency and the Self

08/24/2007

Philosophical Psychology Vol 19 No 3 June 2006

Author uses Dennett's arguments against the Cartesian Theatre as a starting point for a discussion on the self and other concepts of a centralized self-identity. Dennett is hostile to the idea of a unified location or 'brain pearl' that has all systems of the brain in front of it. He uses the analogy of self-organizing systems that give the appearance of centralized intelligence but in fact have none (e.g. termite colonies). The origin of our thinking we have a centralized 'theatre' is our use of words to represent our actions to others-- a useful fiction. (pg 346-7)

Author agrees that there isn't a Cartesian Theatre, but thinks that doesn't mean we end up as termite colonies. Author uses an example of a ship that guides itself by using an internal map. Sensors receive input from the environment. The input is processed using various modules, and a program is run that takes the results of this processed information from all the sources and 'deliberates' about where it is on the map, and what course to set. This could all be displayed graphically, or it could simply be an internal, distributed program. The point is that there is a 'stream' that runs through the 'Joycean Machine', or a program that tries to place itself as a self-representation and 'deliberates' about what course to set. Author considers this the alternative to the self-organizing only model and the Cartesian Theatre model. (pg 349-51)

Author concedes that Dennett does not always talk as eliminatively as that. The tension within Dennett when he seems to endorse a limited 'Joycean Stream' but other times when he insists on only distributed self-organizing systems is reconciled, Author claims, if we take 'language as rooted in the development of explicit self-representation ... representation of ourselves and our states in a causally structured world' (pg 353) For Author, it is this need in social life for self-representation that lead to the Joycean Stream.

Author talks about three types of unity that the Joycean Machine enables:

-Synthetic Unity: the integrating of various disparate information sources.

-Univocity: when the information is integrated into a coherent stream, they are given a 'collective voice'. Here author uses much analogy: like a state-wide referendum that takes all different perspectives and makes them into a 'yes' or 'no', the Joycean Machine is the mouth-piece for a group that has a distributed identity. There is no 'commander' other than the reporter. (pg 356-7)

-Dynamical Unity: The Joycean Machine mediates interactions with other systems, as changes occur.

The important point of all of this is that the reporting of self doesn't mean there is an entity inside the brain 'the self'. The 'reporting' is more like asserting-- a performative that makes it true by concluding what is going on within itself.

8/17/07

Menary, Richard - Attacking the Bounds of Cognition

08/17/2007

Philosophical Psychology Vol 19 No 3 June 2006

Author is undertaking to defend the hypothesis of extended cognition (HEC) and also what author considers is a more radical project that he calls 'cognitive integration', which takes internal (biological) and external vehicles to be integrated into a whole, which is properly considered cognition. The aim of this paper isn't to establish HEC or cognitive integration, but to defend it from the attacks of Adams, Aizawa, and Rupert (A&A).

Author lays out what the cognitive integrationist is committed to:

1) Manipulation thesis: place the 'cognizer' into an environment; agents complete cognitive tasks often by manipulating features of the environment. There are three types of manipulation:
A) Biological cases of coupling (pg 331)
B) Using the environment directly, without representing
C) Manipulation of the external representational system in accordance with cognitive norms

2) Hybrid Mind thesis: cognition is understood as a hybrid process of internal and external systems.

3) Transformation thesis: our cognitive capacities have grown, been transformed, or otherwise augmented by our ability to manipulate, use hybrid processes, and so on.

4) Cognitive Norm thesis: we are able to manipulate external vehicles of cognition because we learn norms that operate on how to manipulate those vehicles. (These norms of external vehicle manipulation are just as cognitive as internal ones.)

A&A, as 'traditional cognitive internalists', do not deny that we use e.g. mathematical symbols to complete cognitive tasks, they just deny such use constitutes a cognitive process. Author claims their objections misconstrue the manipulation thesis and attack a 'weak' parity principle.

The Parity Principle: if an external process were located in the skull, we'd call it cognitive (pg 333) This is supposed to be intuitive, not necessarily an argument for HEC.

A&A's first argument says that if a cognitive process uses/is coupled to object X, it doesn't follow that X is part of the cognition. Author replies that this misunderstands the where/how the cognition is being done. The cognitive integrationist instead has it that cognition is happening with internal processes and objects, together making up cognition. Thus: X is the manipulation of (e.g. the notebook) reciprocally coupled to Y (the brain process) which together constitute the cognitive process of (e.g. remembering). If this seems question begging, author claims that HEC has been independently established, and is beyond the scope of this paper (pg 334).

A&A have a 'intrinsic content' condition that author next attacks. The intrinsic condition seems to be that a process can be counted as cognitive only if it involves at least some intrinsic/non-derived content. Thus a process that involves no intrinsic content is non-cognitive. Somehow mental representations of 'natural objects' are fixed by 'naturalistic conditions on meaning' (Fodor or Millikan or Dretske), and A&A argue that artifical objects can be fixed the same way. The problem here, author claims, is that when you avoid saying that an internalist idea of an artificial object is fixed by conventional content, you stop yourself by using the convential norms that govern use of that artificial object in cognition. But we do use these norms in manipulation of these artificial objects. So either the objection takes us to be less competent than we are, or the objection posits intrinsic content that is suspiciously similar to convential content. (For an in-depth review of the dialectic, see pg 334-7)

A&A object that we have no good way of making a science out of the combination of brains and external tools, since external tools are all so disparate. A related objection from Rupert is that notebooks and any external tool you can use so far can't really be used when keeping up in conversation, so conversational memory doesn't work if it is external. (pg 339) Author replies to A&A by saying that they miss the entire point. It isn't that cognitive integrationists say that what happens externally is just like what happens internally! (pg 340) It is that, instead, the external vehicles take part in a hybrid process of cognition. Author replies to Rupert that he may be right, but other sorts of memory work differently.

8/10/07

Fisher, Justin - Why Nothing Mental Is Just In The Head

08/10/2007

Nous, Vol 41 No 2 2007

This paper uses a counter-example to 'mental internalism' to show that it isn't just what happens 'in the head' that influences mental events. Author defines a 'mental internalist' early:
A Mental Internalist believes that an individual's mental features supervene on what is in that individual's head at that time. Likewise for two individuals with the same mechanical layout: same things inside the head = same mental features. Author explains how some of this has been challenged by 'classical' externalist arguments (Putnam, Kripke, Burge), particularly on the side of the content of mental features (for instance, the content of my thought that 'Water is wet'), and in what justifies a belief. Of course, externalism of this sort has been open to challenge from a 'narrow content' view of the content of beliefs-- but author tries to get away from this. Classical externalist arguments haven't touched many of the hallmark mental features: phenomenal experiences, rationality, moral character, emotions, propositional-attitude-types. Author constructs an example that disproves mental internalism:

Imagine a world where there are 100 radiation 'pulses' per second shooting around. They are disruptive to the human physiology so that our mechanics/mental causation would go haywire if we were on that world 'Pulse-world': we would go quite mad. However, there are 'Pulselings' who have evolved to be just like humans except that their mental mechanics do just fine (maybe even need) to have these pulses going through their heads 100x/second. Now, one Pulseling Paula is having the experience of driving, and an Earthling Edna is having the experience of playing the saxophone. At some point in time t (in between pulses) author stipulates that these two people's mechanical/physical/inside-the-head properties are identical. (pg 321-2) If this is possible, there is a difference between mechanical inside-the-head and mental features. Thus mental internalism is false.

The next section deals with whether this example is possible. Author claims his example rests on three assumptions:
1) Our mental features are produced as the consequence of relatively simple interactions between many elements in our heads
2) These pulses 'coax' the elements in our heads to change the mechanics of how they operate
3) If these pulses change small elements in our heads, they can change large ones too
The moral of the story is that all cognitive systems depend deeply on the appropriate support (or at least non-interference) from their surroundings. (pg 324)

Author next considers replies to his example. The first is the other-minds skeptic. Since nobody can say much to him, author can't either. Nobody can convince the other-minds skeptic that other humans have mental features, let alone Pulselings. Another defense might be that a pulseling that receives these pulses is disqualified from having mental features attributed to her, because of these pulses. This is ad hoc, and denies explanatory power (since it certainly looks as though Pulselings are intelligent, have feelings, and so on), and might also disqualify ourselves as well (since we might be dependent on some sort of environmental factor).

Author considers two possible alternatives to mental internalism. The first is 'wide functionalism', which expands the mental features to some more of the subject's current surroundings. Author dislikes this in favor of 'teleo-functionalist' historical perspective, which takes into account the history of the subject in order to determine what the normal mechanics are for 'in the head' mental features. Author espouses the Principle of Mental Inertia:

--Altering things outside a creature's head won't significantly change the progression of mental states that that creature will undergo, unless those external alterations also bring about change within the creature's head. (pg 329) [What? Things won't be different unless they're different?!?]

Author briefly describes why his teleo-functionalist account is superior to the wide functionalist account, by suggesting that both Edna (Earthling) and Paula (Pulseling) are de-brained and their brains are thrust into identical vats: each would still have the same surroundings but their mental features would be different. Thus wide functionalism would fail here, but the Principle of Mental Inertia would be consistent with this result.

8/2/07

Montero, Barbara - Physicalism Could Be True Even If Mary Learns Something New

08/03/2007

The Philosophical Quarterly, Vol 57 No 227 April 2007

In this paper the thesis is that Mary would lack the concept of 'what it is like to see red', even if she knew what happened on the lower-level physical level, and could deduce what would happen on the higher-level physical level. Author dubs this the 'missing-concept' reply to the knowledge argument.

Author starts by discussing a 'less than ideal' knowledge argument that is open to flaws. She uses this as a starting point for some of her claims as replies. The less than ideal argument starts with 'Mary knows all the facts of physics, chemistry and neurophysiology...'. This is open to problems because there may be other physical facts that aren't included in these fields. There could be 'higher-level' physical facts (those that constitute/determine the experience of red) that aren't, strictly speaking, included in physics, chemistry or neuropsych. This is consistent with what author calls the 'non-reductive' physicalist position, with a conception of the 'broadly physical'. The 'broadly physical' is that mental facts are physical facts, whatever those facts may be (pg 179).

This leads to a discussion of what it is to be physical at all. Author begins with saying that as long as a property is either fundamental and physical or determined by fundamental physical properties, it is broadly physical. Much talk in the sciences involves talk of deducing higher-level physical facts from lower-level ones. There should be no reason why, in principle, this can't be done. This is the case, author points out, only if all fundamental physical facts are taken to be 'structural/relational' facts. If we construe the physical as the 'non-mental', then we won't have this necessary connection. (pg 181-2) [Doesn't this beg the question?] Only on a certain understanding of the physical as being ultimately accessible to physics using structure, position, charge, etc. can higher-level properties be deducible from lower-level physical ones.

The fixed Mary argument takes Mary to know all the fundamental lower-level physical facts and have perfect reasoning and deduction skills. Author abandons the previous argument she uses (above) and agrees that all higher-level facts are deducible from lower-level ones. Presumably, this can be done a priori. However, can it be done without the relevant concepts? One might think that this is what a priori just means. However, author claims that a priori means that the truth of the conclusion is justified from the truth of the premises without reference to empirical studies. This doesn't mean the conclusion can be reached by simply looking at the premises-- sometimes you'd need the relevant concepts to employ. (pg 183-87) Presumably, Mary could infer "Ahh, seeing red would look like this", except that she wouldn't understand that 'this' refers to, since she lacked the relevant concept of 'the experience of seeing red'.

The last bit of the paper tries to show that author's reply to the Mary argument is different from the 'old fact, new presentation' reply. The 'old fact, new presentation' argument uses identity between (brain-state B) and (seeing red). The 'non-reductive physicalist', however need not hold this identity-- in the sense that the two propositions have the same truth-value. (pg 188) [WHAT?!?]