Loading ...
Sorry, an error occurred while loading the content.

Apriori nature of mind/body problem

Expand Messages
  • ilivetothink
    Hi all, No idea if anyone here will be interested, but perhaps if you think there is some possibilty of constructing a computer which could sustain
    Message 1 of 4 , Nov 4, 2007
    • 0 Attachment
      Hi all,

      No idea if anyone here will be interested, but perhaps if you think
      there is some possibilty of constructing a computer which could
      sustain consicousness, this will point out some difficulties...

      would be very interested in any
      questions/comments/criticisms/nitpicks...

      First section defines Materialism, second explains the apriori nature
      of the explanatory gap, third is Levine's critique of apriority,
      fourth is my critique of Levine.

      Cheers,
      Kelvin






      Introduction

      Joseph Levine is a Materialist. In other words, he believes
      that "Only non-mental properties are instantiated in a basic way; all
      mental properties are instantiated by being realized by the
      instantiation of other, non-mental properties" (p21). By "basic"
      Levine means "irreducible", and so the basic properties (and
      entities) are the ones that will be mentioned in a completed
      fundamental physics. `Realization', for Levine, is a relation between
      properties, distinct from the relations of identity and of lawful
      correlation. He defines it as follows: "The instantiation of property
      is A is realized by the instantiation of property B just in case the
      very fact alone of B's instantiation constitutes the instantiation of
      A [but not vice versa]" (p13). I have added "but not vice versa" into
      Levine's definition to highlight the crucial distinction between
      realization and identity. Levine's favourite examples of realization
      come from computer science, where functional properties, such as
      computer programs, are physically realized by the electronic
      mechanisms in the computer. Consider a chess program. The reason the
      relation between the program and the electronic mechanisms that
      sustain it is not one of identity is because a chess program could be
      realised by something else, such as a person moving around chess
      pieces on a chessboard. The same holds for mental properties such as
      intelligence or pain. We might say that the firing of C-fibers
      realizes pain in human beings, but in other creatures that feel pain,
      it may be something quite different.
      If Materialism is true then we should be able to explain how
      mental properties are realized by non-mental properties, or in other
      words, how mental properties reduce to non-mental properties. Levine
      believes we are well on our way in doing this for rationality and
      intentionality, but he thinks there is a major problem for conscious
      experience, or more technically, phenomenal properties. As defined by
      Thomas Nagel, a property is phenomenal just in case there is
      something it is like to have that property. So pain is a phenomenal
      property because there is something it is like to feel pain. Levine
      thinks that the greatest challenge for Materialism is the so-
      called "conceivability argument", particularly David Chalmers' modern
      defense of it. This argument infers an ontological gap between
      consciousness and non-mental physical properties, from an explanatory
      gap, that is, from a theory explaining why we cannot reductively
      explain consciousness in non-mental physical terms. While Levine is
      very sympathetic with the explanatory gap (he actually invented the
      notion 24 years ago), he refuses to accept an ontological gap.
      Chalmers' explanatory gap theory is set up in a way that makes it
      easy for him to infer an ontological gap. Thus, Levine criticizes the
      semantic assumptions involved in Chalmers' theory and constructs a
      quite different semantics of his own. In this paper, I will argue
      that Levine's philosophy of language cannot work and so neither can
      his arguments against Chalmers' from which I will conclude that
      Levine has failed to save Materialism from the threat of the
      conceivability argument.

      The object of Levine's attack: Chalmers' explanatory gap theory

      Let's begin by looking at Chalmers' explanatory gap theory. As
      mentioned, Materialism requires reductive analyses, because it does
      not want to treat phenomenal properties as basic – they must be
      reducible to their physical realizers. But for Chalmers, reductive
      explanations involve two specific requirements, which reductions of
      phenomenal properties can't seem to meet, that is, conceptual
      analysis and scientific explanation. I shall illustrate with the
      simple example of the reduction of water to H20:

      [Conceptual Analysis:] Water is what plays the water role (Apriori)
      [Scientific Explanation:] H20 is what plays the water role
      (Aposteriori)
      [Reductive Conclusion:] Water is (identical to) H20.

      The first premise consists in cashing out our concept of water, that
      is, what we mean when we speak of water, so that we are clear on what
      we want to reductively explain. Chalmers takes natural kind concepts
      like "water" to be functional or causal role concepts, and so, a
      complete conceptual analysis will tell us what the water role is, and
      that water is what plays that role. A rough preliminary conceptual
      analysis might tell us that water is the dominant clear, drinkable
      liquid in our seas, rivers, lakes and rains. Given that conceptual
      analysis tells us about what we already know tacitly in virtue of
      linguistic competence, this truth about water is known apriori.
      Levine disagrees sharply with this, and so I shall come back to
      conceptual analysis in detail later.
      Concerning the second premise, this is where scientists tell us that
      H20 is the dominant clear, drinkable liquid in our seas, rivers,
      lakes and rains. As we shall soon see, for Chalmers' argument against
      Materialism to work, Chalmers needs it to be the case that we can (at
      least in principle) be told that H20 plays this functional role
      purely in microphysical terms. So for example, a complete
      microphysical theory of the universe will tell us the geometrical
      properties of H20 molecules in space-time, from which, given
      possession of the concept of liquidity, we should be able to infer
      that H20 molecules exemplify liquidity. It will also tell us about
      spatial locations of H20, from which we should be able to infer,
      given possession of the concepts of lakes and rains that H20
      molecules are in our lakes and fall from our skies. We thus get told
      in terms neutral to macrophysical and microphysical vocabulary (e.g.
      in mathematical, geometrical and causal terms), that H20 plays the
      water role, from which we can infer, given our apriori premise, that
      water is H20.
      For Chalmers, unless we can do this for phenomenal properties, then
      we are not going to get a reduction, and we will have an explanatory
      gap. In that case, Materialism may still be true, we just won't be
      able to establish it. So Materialism needs something like:

      [Conceptual Analysis:] Pain is what plays the pain role (Apriori???)
      [Scientific Explanation:] C-Fibers firing is what (at least in
      humans) plays the pain role (Aposteriori)
      [Reductive Conclusion:] Pain is (realised in humans by) the firing of
      C-fibers

      Perhaps premise one tells us that pain is what typically causes
      certain behaviours like writhing while being caused by certain inputs
      such as tissue damage. The problem here, for Chalmers, is that this
      truth about pain is not apriori. Chalmers backs this up by saying
      that whatever constitutes our concept of pain, it is constituted in
      such a way that we can conceive of the instantiation of any such
      functions in the absence of pain. For example we can conceive of
      creatures who constantly instantiate this so-called pain role but
      never feel pain. The same cannot be said for non-phenomenal natural
      kind terms. Thus, phenomenal concepts, for Chalmers, are unique in
      that they are not constituted by any apriori functional role
      specification, implying that the scientific explanation cannot get
      off the ground, because there is nothing there for it to explain in
      functional terms. This is his explanatory gap theory.

      Chalmers then uses this unique property of phenomenal
      concepts in his argument against Materialism:

      (1) P&~Q is conceivable (cannot be ruled out apriori)
      (2) If P&~Q is conceivable then P&~Q is 1-possible
      (3) If P&~Q is 1-possible then either P&~Q is 2-possible or Russelian
      Monism is true
      (4) If P&~Q is 2-possible then Materialism is false
      Therefore, either Materialism is false or Russelian Monism is true

      Q is a phenomenal truth, e.g., that someone is consciously
      experiencing pain, while P is the conjunction of all microphysical
      truths about the universe. P&~Q thus specifies a world identical to
      ours in all microphysical respects, but which differs phenomenally.
      We can see that premise one is substantive once we see that if we
      replace Q with any non-phenomenal macroscopic truth, say, a truth
      about water (=W), then P&~W will be inconceivable. That is, P will
      tell us in topic neutral terms what plays the water role and all the
      truths about the stuff that plays the water role. Thus, if we have
      the concept of water, then we can infer apriori, from P, W. We could
      thus rule P&~W out apriori. P&~Q cannot be ruled out apriori
      precisely because the phenomenal term in Q has no apriori functional
      role. Unfortunately I don't have time to explain this argument I just
      present it show what's on the line, I will, however, be happy to
      answer any questions concerning it.

      Levine's way out of the argument

      To escape the conclusion of the argument, Levine attempts to
      trivialize the first premise so much so that nothing substantial can
      be inferred from it. In particular, he wants to show that premise one
      is true, not because of anything that is unique to phenomenal
      concepts as Chalmers argues, but because practically nothing can be
      ruled out apriori, due to the scarcity of apriori truths. In
      particular, Levine wants to show that P&~W cannot be ruled out
      apriori for the very same reason that P&~Q cannot be ruled out
      apriori. Levine's conclusion then, is as follows: Clearly, one would
      not conclude that there is an ontological gap between H20 and water
      from the premise that P&~W can be ruled out apriori. So to one should
      not conclude that there is an ontological gap between microphysics
      and phenomenal properties, from the premise that P&~Q cannot be ruled
      out apriori.
      To clarify the disagreement over the extent of the set of apriori
      truths, Levine introduces two concepts, `semantic form' and `logical
      form'. For Levine (apart from some specific examples that I'll
      mention later), the apriori truths are just the logical truths and so
      he uses the term "logical form" to name the source of this apriori
      knowledge. But despite the importance of this notion, he says nothing
      about it, except that it is what allows us to know that logical
      contradictions cannot express truths. To distinguish himself from
      Chalmers, he notes that Chalmers believes that `semantic form' is
      also a source of apriori knowledge. Levine is much clearer on
      semantic form and defines it as:

      "…information concerning the application conditions of concepts. So
      if I know that all bachelors are unmarried a priori it's because I
      know that nothing counts as a bachelor unless its unmarried" (p41).

      So the semantic form for natural kind terms such as "water" consist
      in their functional role specifications. With these concepts in hand,
      Levine states the difference between himself and Chalmers, while
      Chalmers believes that semantic form is a source of apriori
      knowledge, Levine doesn't, Levine believes that the primary source is
      logical form. If Levine is right, then even if P tells us that H20
      plays the water role, we cannot then infer apriori, solely through
      possessing the concept of water, that water is H20, for the water
      role is not connected apriori to our term "Water". And so, if P&~Q is
      conceivable in this sense, then nothing substantial can be concluded
      from it. The question now becomes, why should we think that semantic
      form is not a source of apriori knowledge?

      How does Levine Restrict apriority to just logical form?

      Levine's main tactic is to show that we have no reason to assume that
      the class of apriori truths consists in anything more than the
      logical truths. I will present what I take to be the two most
      significant elements of Levine's attack. Firstly, his attempt to show
      that the theoretical tool that Chalmers uses to uncover apriori
      truths – conceptual analysis – is unfit for the task. And secondly,
      his appeal to the causal theory of reference, which, in order to have
      a coherent philosophy of language, is needed to supplement his
      scepticism about the apriori.

      He believes that there is no reason to think that the results of
      conceptual analysis are apriori

      When we do conceptual analysis, we give descriptions of all sorts of
      possible worlds not involving a term T, ask ourselves what T refers
      to or when T should be applied in these worlds, and then generalise
      over our answers. Such generalizations are meant to capture general
      facts about the application conditions, or meanings of T. Sentences
      that describe such general facts are meant to be apriori, because
      they survive the method of possible cases. That is, if a conceptual
      analysis of our term "water" tells us that water is what plays the
      water role, then we can test whether or not this sentence is apriori,
      by trying stipulate a possible case in which we would consider it to
      be false of water. If we cannot, then that is good evidence that the
      truth of the sentence is not dependent on the way the world turns, so
      it is not aposteriori, but it is dependent on how we wish to apply
      the term in these worlds, which is something we know apriori.
      Perhaps the most famous example, that is considered to be conceptual
      analysis at least by defenders of conceptual analysis, is the
      analysis of "water" given by Hilary Putnam . Putnam imagined that
      somewhere in the galaxy there exists "twin-Earth" which is identical
      to Earth, except that the transparent liquid in its seas rivers lakes
      and rains (i.e. what plays its water role) is XYZ rather than H20.
      Putnam then asked us if our word "water" would refer to the stuff
      that plays their water role as well as what plays ours. Most people's
      intuitions tell them that it doesn't, that the watery stuff on their
      planet is not H20 and so is not water, it just seems like water.
      Let's reformulate the example. Now imagine a possible world where we
      find out that we were wrong in thinking that what plays our water
      role is H20, as it is actually XYZ (so in a sense we find out that
      twin-Earth is our earth). Would we say that "water" does not refer in
      this possible world, or that it does, as it refers to XYZ? The
      typical response is that it refers to XYZ.
      What do these modal intuitions – intuitions about how to apply terms
      in imagined cases – show us? They seem to show us what counts as
      water. What counts as water is something like the transparent liquid
      stuff that's in our seas rivers lakes and rains. Considering twin-
      Earth as a planet in a different galaxy to ours, as Putnam invited us
      to, brought out the indexical element – "our". And also that "water"
      refers to the occupant of the water role and not the role property
      itself (so that water is not identical to the water role but to what
      plays the water role). And considering twin-Earth as a way our planet
      might turn out to be shows us which associated properties – e.g.
      transparency, liquidity, being in lakes and rivers etc. – go into
      the `water role' specification that we use to pick out the water role
      occupant.
      Levine denies that such results represents apriori knowledge.
      Concerning what the water role consists in, Levine believes that
      these are merely our strongest beliefs about water, which are
      empirical. And so, we use these empirical beliefs, which may well be
      false of water, to pick out the extension of water in worlds.
      Concerning the identity of water with the role occupant rather than
      the role, Levine admits that empirical enquiry cannot help and that
      only modal intuitions can inform us of this. But he argues that
      whatever our concepts consist in, what constitutes them cannot be a
      capacity for conceptual analysis, that is, this ability to reflect on
      our use of concepts in possible worlds. This is because higher
      animals have concepts, but cannot perform conceptual analysis.
      Therefore, there is no reason to think that conceptual analysis tells
      us about the requirements for the possession of concepts, be it
      apriori knowledge or whatever. I'll come back to this.

      He appeals to a Causal theory of reference

      A theory of reference is one that answers the following question: Why
      do our natural kind terms have the referents that they have? Chalmers
      has an answer that invokes his broad conception of apriority:
      Firstly, when speakers become competent with the use of the term,
      they tacitly associate with it a functional role definition – water
      is what plays the water role. Secondly, the world must play its part –
      it must provide a referent that plays that functional role. Thus,
      reference is determined by an apriori associated functional role and
      the feature (if any) of the world that plays that role. Accordingly,
      coming to know what a term refers to will consist in two parts – an
      apriori and an aposteriori part. To know what "water" refers to, one
      must become competent with the conventional use of "water", that is,
      one must know that water is what plays the water role, and then, one
      must empirically discover what actually plays the water role. The
      former is apriori knowledge the latter aposteriori.
      Levine needs to answer the question without appeal to apriority. This
      is because, in taking the water role to be aposteriori, he must admit
      that the water role specification might be false of water. In that
      case, of course (possibly) false associated statements about water
      cannot explain why "water" refers to water. So Levine instead appeals
      to causal theories of reference.
      Levine does not provide his own theory, all he says is that "What
      determines extension […] are the external relations that obtain
      between a representation and what it represents. On most theories the
      relevant relation is something like causal covariation" (p54).
      Nonetheless, we can illustrate what a theory of this form looks like
      with Michael Devitt and Kim Sterelny's causal-historical theory. This
      theory is developed out of the insights of Kripke and Putnam, which
      is where these causal theories began. For D&S, a name or natural kind
      term gets introduced in the presence of an object, or paradigmatic
      instance or sample of some natural kind, which establishes a causal
      link between the term and its extension. Those who introduce the term
      go on to spread the term around the community, making more causal
      chains, which all ultimately lead back to the dubbing ceremony and so
      to the extension. What determines the reference of a name when I use
      it is the causal chain that leads down to its referent. What
      determines the reference of a natural kind term when I use it is the
      causal chain which leads down to the paradigm instance(s) or sample
      (s) of the kind which shares an underlying nature or sameness
      relation to all other instances or samples of the term's extension.
      These causal relations are external to the mind and are therefore
      knowable only aposteriori. In other words, knowledge of such
      reference determiners have not been drilled into the mind through the
      practice of language acquisition at all such that we can tease out
      info about them by conceptual analysis. Apriority thus plays no
      significant role, for the causal theorist, in determining reference.

      My Critique of Levine

      Concerning the alternative theory of reference, I will argue that the
      causal theory suffers from a problem that can only be resolved by
      invoking a broad conception of apriority. Concerning conceptual
      analysis, I will argue that Levine misconstrues it, and that once we
      get it clear and use it to illuminate why logical truths might be
      apriori as Levine assumes they are, we will see that Levine's
      conception of the extent of apriority is arbitrary, while Chalmers'
      isn't.

      Rejection of the Causal Theory (the qua-problem)

      Consider the dubbing ceremony for "water". What makes it the case
      that the term refers to water qua natural kind as opposed to, say,
      just the lake involved in the dubbing ceremony, so that "water" is
      really a proper name? Or to illustrate the objection in another way;
      what makes it the case that the term refers to water qua substance as
      opposed to, say, the colour or weight of the sample water? So, on the
      causal account, why doesn't "water" refer to an individual lake, or
      to a particular colour or to a particular weight? Appeal to causal
      links cannot solve this problem. It seems the only way out of the qua-
      problem, is to leak some apriori knowledge in. So we know apriori
      that water is a natural kind. But this isn't enough. What now
      stops "water" from referring to, say, anything physical? Looks like
      we will have to leak some more apriori knowledge in – perhaps that it
      is a liquid. But this still won't do, for then why doesn't the term
      refer to all kinds of liquid? It seems we need to say that it refers
      to this kind of liquid while pointing at, and having in mind, say,
      paradigmatic instances of our seas, rivers, lakes and rains.
      What should we conclude from this about the explanatory worth of the
      causal theory of reference? On the face of it, it seems that if we
      keep hitting Levine with these qua-problems, we shall inevitably
      arrive at Chalmers' theory of the apriori. By "water" we mean the
      stuff that plays the water role, and so, we only need to reflect on
      our concepts, without appeal to empirical information, to know that
      water is what plays the water role. And once we are at this stage, it
      seems as though we have enough to determine reference, for we have an
      apriori water role specification, and then we just leave it up to the
      world to provide a role occupant.

      Clarifying Conceptual Analysis and Drawing the apriori/aposteriori
      divide

      Levine gives a particular interpretation to conceptual analysis. As
      he says in a crucial footnote, Chalmers "is arguing from an analysis
      of what it is to possess a concept to the claim that one must have a
      priori knowledge of how extension is determined by context. It is
      this inference that is being rejected" (p186). But what is
      this "analysis of what it is to possess a concept" that Chalmers is
      allegedly arguing from? Levine thinks that Chalmers is saying that
      the ability to reflect on our practice with conceptual analysis is
      what constitutes, or is at least necessary for, our linguistic
      practice.
      However, the ability to perform conceptual analysis does not
      constitute our apriori knowledge, but rather, it makes explicit our
      apriori knowledge. So what then, is the difference between we who are
      capable of conceptual analysis, and other concept possessors, or more
      specifically, other possessors of apriori knowledge, such as higher
      animals? The difference is in the capacity to do conceptual analysis
      of course, and this capacity involves more than the possession of
      concepts – which is precisely what Levine has overlooked. To do
      conceptual analysis, you must possess two other capacities: firstly,
      a powerful imaginative capacity, and secondly, the ability to
      describe complex imaginings with a complex linguistic system. Only
      then can you describe complex possible worlds before asking yourself
      what your concepts designate in those worlds. Without the latter two
      capacities, the apriori knowledge that constitutes your concepts will
      forever remain silent – trapped deep within the mind, as is the case
      with higher animals.
      With our methodology for uncovering apriori truths clear, we
      should now turn to Levine's distinction between logical form and
      semantic form. Levine expresses perplexity as to the nature of
      logical form, that is, the source of our apriori knowledge of logical
      truths, as he says:

      "I know a priori that I cannot both quit and not quit my job, because
      I know that no statement of this form can express a truth; it's a
      logical contradiction. How do I know that logical contradictions
      cannot express truths? This is a deep question that I cannot take up
      here" (p41).

      I believe that the source of our apriori knowledge of logical truths,
      that is, logical form, simply consists in the semantic form of
      logical terms. I am suggesting that we know that a logical
      contradiction, say of the form "X and not X" cannot express a truth
      due to our knowledge of the application conditions of logical terms.
      Levine does not explain what he means by, "application
      conditions", but read literally, the expression just means
      information about the correct conditions under which we should apply
      (or use) a term. And all terms, logical and non-logical alike, have
      these. For all such terms, we have an intuitive understanding of the
      conditions under which they ought to be used. So what might the
      application conditions of logical terms be? One suggestion might be
      that their truth tables capture their application conditions. And
      there are two significant similarities between this and the
      suggestion that "water"'s application conditions are captured by the
      water role specification.
      Firstly, while one must have a tacit understanding of the water role
      to be able to use "water" with the community, similarly, it seems a
      tacit understanding of the truth tables is necessary for one to be
      able to apply logical terms correctly. Secondly, how did we come up
      with truth tables? We did it the same way we came up with the water
      role specification – conceptual analysis. Truth tables are idealized
      generalizations over our actual use of logical connectives (in
      worlds, if you like) that hold between propositions. This is why
      propositions are replaced by variables within the tables. And most
      importantly, the truth tables are correct if they survive the method
      of possible cases. If you can imagine a situation that could be truly
      described by a complex proposition whose truth is not predicted by
      the truth tables of the relevant logical connectives, then those
      truth tables will have been ruled out by conceptual analysis.
      And so I can see no reason why we should be privileging logical terms
      over non-logical terms when it comes to determining the extent of
      apriori truths. Our methods for establishing that a logical truth is
      apriori, is just conceptual analysis, and so there is no reason why
      we cannot perform conceptual analysis on a sentence as a whole,
      rather than just arbitrarily privileging its logical terms. Just to
      drive the arbitrariness objection even further, Levine actually
      thinks that the class of apriori truths is slightly bigger than the
      set of logical truths. After his discussion of the modal intuition
      that water is a role occupant rather than a role property, he says:

      "Finally it is open […] to admit that on the question of what kind of
      thing water is—a role-property or a substance occupying the role—we
      do have a priori access to the answer. But that isn't sufficient to
      render [P&~W] conceptually impossible, which is the point at issue
      here. So long as the specification of the role itself is not known a
      priori, statement [P&~W] cannot be ruled out on a priori grounds"
      (p62).

      But it is not a logical truth that water is an occupant property. So
      then what else could make it true if not the fact that the
      application conditions of "water" disallow us to be able to conceive
      a world where water is a role property? So there seems to be no
      substantial difference between this and the suggestion that the
      application conditions of logical terms, as captured in their truth
      tables, disallow us to conceive of a world where one can both quite
      and not quit their job.
      Levine also says:

      "Aristotle […] could not have failed to be human. Nothing that is not
      human counts as the same individual as Aristotle […] it must be by
      virtue of an a priori principle to the effect that "human being" is a
      privileged sortal. That is, if the predicate applies at all to an
      object, then it applies to that object in every possible world in
      which the object exists" (p43).

      Once again, the sentence "if Aristotle is human then nothing that's
      not human could be Aristotle", is not a logical truth. How then do we
      know apriori that it is true? Surely by means of knowing the apriori
      principle expressed the last sentence of the quote, which surely
      captures an aspect of the application conditions which govern the use
      of "human being" uncovered through conceptual analysis.
      We can now pass this reasoning over to the crucial
      sentence "water is what plays the water role". Is it really the case
      that no empirical evidence could ever show us that water is not
      actually the dominant clear, drinkable liquid in our seas, rivers,
      lakes and rains? I can't see how. I can certainly imagine finding out
      that water is not in the lakes, but that would just be to discover
      that the dominant clear, drinkable liquid in our seas, rivers and
      rains is not in our lakes. So the real apriori truth could be thought
      of as:

      "Water is the stuff that exemplifies most of the basic properties
      that I use to pick out its instances, such as being clear, drinkable,
      and being the dominant stuff in our seas, rivers, lakes and rains".

      I see no reason to think that this is aposteriori, it seems clear to
      me that, or at least something like it, exhausts our concept of
      water.
      And so, "water" does have a functional role specification, which can
      be completely cashed in the above manner. If Chalmers is right that
      this is required to give a reductive analysis of water, where science
      tells us what more fundamental property plays this functional role,
      then it seems we do have a significant difference between phenomenal
      terms and ordinary natural kind terms. We can conceive of pain
      obtaining independently of the instantiation of any functional role
      property. Hence, no functional role specification can capture our
      concept of pain. And if scientific reductions do require such
      specifications, then we are left with an explanatory gap between
      phenomenal properties and the physical world. I hope I have shown
      that Levine's defense of Materialism has failed, and that if we wish
      to escape the conceivability argument, then this is not the way.
    • Ron
      I was not sure if you wanted the machine to be conscious or if you wanted it to be linked to people to help them to be conscious. It would seem to me that a
      Message 2 of 4 , Nov 7, 2007
      • 0 Attachment
        I was not sure if you wanted the machine to be conscious or if you
        wanted it to be linked to people to help them to be conscious.

        It would seem to me that a generalized gaussian representation of
        others is the father of an generalized gaussian representation of
        self. Seeing your self then would be defined as a conscious
        experience.

        Ron


        --- In artificialintelligencegroup@yahoogroups.com, "ilivetothink"
        <mcqke464@...> wrote:
        >
        > Hi all,
        >
        > No idea if anyone here will be interested, but perhaps if you
        think
        > there is some possibilty of constructing a computer which could
        > sustain consicousness, this will point out some difficulties...
        >
        > would be very interested in any
        > questions/comments/criticisms/nitpicks...
        >
        > First section defines Materialism, second explains the apriori
        nature
        > of the explanatory gap, third is Levine's critique of apriority,
        > fourth is my critique of Levine.
        >
        > Cheers,
        > Kelvin
        >
        >
        >
        >
        >
        >
        > Introduction
        >
        > Joseph Levine is a Materialist. In other words, he believes
        > that "Only non-mental properties are instantiated in a basic way;
        all
        > mental properties are instantiated by being realized by the
        > instantiation of other, non-mental properties" (p21). By "basic"
        > Levine means "irreducible", and so the basic properties (and
        > entities) are the ones that will be mentioned in a completed
        > fundamental physics. `Realization', for Levine, is a relation
        between
        > properties, distinct from the relations of identity and of lawful
        > correlation. He defines it as follows: "The instantiation of
        property
        > is A is realized by the instantiation of property B just in case
        the
        > very fact alone of B's instantiation constitutes the instantiation
        of
        > A [but not vice versa]" (p13). I have added "but not vice versa"
        into
        > Levine's definition to highlight the crucial distinction between
        > realization and identity. Levine's favourite examples of
        realization
        > come from computer science, where functional properties, such as
        > computer programs, are physically realized by the electronic
        > mechanisms in the computer. Consider a chess program. The reason
        the
        > relation between the program and the electronic mechanisms that
        > sustain it is not one of identity is because a chess program could
        be
        > realised by something else, such as a person moving around chess
        > pieces on a chessboard. The same holds for mental properties such
        as
        > intelligence or pain. We might say that the firing of C-fibers
        > realizes pain in human beings, but in other creatures that feel
        pain,
        > it may be something quite different.
        > If Materialism is true then we should be able to explain how
        > mental properties are realized by non-mental properties, or in
        other
        > words, how mental properties reduce to non-mental properties.
        Levine
        > believes we are well on our way in doing this for rationality and
        > intentionality, but he thinks there is a major problem for
        conscious
        > experience, or more technically, phenomenal properties. As defined
        by
        > Thomas Nagel, a property is phenomenal just in case there is
        > something it is like to have that property. So pain is a
        phenomenal
        > property because there is something it is like to feel pain.
        Levine
        > thinks that the greatest challenge for Materialism is the so-
        > called "conceivability argument", particularly David Chalmers'
        modern
        > defense of it. This argument infers an ontological gap between
        > consciousness and non-mental physical properties, from an
        explanatory
        > gap, that is, from a theory explaining why we cannot reductively
        > explain consciousness in non-mental physical terms. While Levine
        is
        > very sympathetic with the explanatory gap (he actually invented
        the
        > notion 24 years ago), he refuses to accept an ontological gap.
        > Chalmers' explanatory gap theory is set up in a way that makes it
        > easy for him to infer an ontological gap. Thus, Levine criticizes
        the
        > semantic assumptions involved in Chalmers' theory and constructs a
        > quite different semantics of his own. In this paper, I will argue
        > that Levine's philosophy of language cannot work and so neither
        can
        > his arguments against Chalmers' from which I will conclude that
        > Levine has failed to save Materialism from the threat of the
        > conceivability argument.
        >
        > The object of Levine's attack: Chalmers' explanatory gap theory
        >
        > Let's begin by looking at Chalmers' explanatory gap theory. As
        > mentioned, Materialism requires reductive analyses, because it
        does
        > not want to treat phenomenal properties as basic – they must be
        > reducible to their physical realizers. But for Chalmers, reductive
        > explanations involve two specific requirements, which reductions
        of
        > phenomenal properties can't seem to meet, that is, conceptual
        > analysis and scientific explanation. I shall illustrate with the
        > simple example of the reduction of water to H20:
        >
        > [Conceptual Analysis:] Water is what plays the water role (Apriori)
        > [Scientific Explanation:] H20 is what plays the water role
        > (Aposteriori)
        > [Reductive Conclusion:] Water is (identical to) H20.
        >
        > The first premise consists in cashing out our concept of water,
        that
        > is, what we mean when we speak of water, so that we are clear on
        what
        > we want to reductively explain. Chalmers takes natural kind
        concepts
        > like "water" to be functional or causal role concepts, and so, a
        > complete conceptual analysis will tell us what the water role is,
        and
        > that water is what plays that role. A rough preliminary conceptual
        > analysis might tell us that water is the dominant clear, drinkable
        > liquid in our seas, rivers, lakes and rains. Given that conceptual
        > analysis tells us about what we already know tacitly in virtue of
        > linguistic competence, this truth about water is known apriori.
        > Levine disagrees sharply with this, and so I shall come back to
        > conceptual analysis in detail later.
        > Concerning the second premise, this is where scientists tell us
        that
        > H20 is the dominant clear, drinkable liquid in our seas, rivers,
        > lakes and rains. As we shall soon see, for Chalmers' argument
        against
        > Materialism to work, Chalmers needs it to be the case that we can
        (at
        > least in principle) be told that H20 plays this functional role
        > purely in microphysical terms. So for example, a complete
        > microphysical theory of the universe will tell us the geometrical
        > properties of H20 molecules in space-time, from which, given
        > possession of the concept of liquidity, we should be able to infer
        > that H20 molecules exemplify liquidity. It will also tell us about
        > spatial locations of H20, from which we should be able to infer,
        > given possession of the concepts of lakes and rains that H20
        > molecules are in our lakes and fall from our skies. We thus get
        told
        > in terms neutral to macrophysical and microphysical vocabulary
        (e.g.
        > in mathematical, geometrical and causal terms), that H20 plays the
        > water role, from which we can infer, given our apriori premise,
        that
        > water is H20.
        > For Chalmers, unless we can do this for phenomenal properties,
        then
        > we are not going to get a reduction, and we will have an
        explanatory
        > gap. In that case, Materialism may still be true, we just won't be
        > able to establish it. So Materialism needs something like:
        >
        > [Conceptual Analysis:] Pain is what plays the pain role
        (Apriori???)
        > [Scientific Explanation:] C-Fibers firing is what (at least in
        > humans) plays the pain role (Aposteriori)
        > [Reductive Conclusion:] Pain is (realised in humans by) the firing
        of
        > C-fibers
        >
        > Perhaps premise one tells us that pain is what typically causes
        > certain behaviours like writhing while being caused by certain
        inputs
        > such as tissue damage. The problem here, for Chalmers, is that
        this
        > truth about pain is not apriori. Chalmers backs this up by saying
        > that whatever constitutes our concept of pain, it is constituted
        in
        > such a way that we can conceive of the instantiation of any such
        > functions in the absence of pain. For example we can conceive of
        > creatures who constantly instantiate this so-called pain role but
        > never feel pain. The same cannot be said for non-phenomenal
        natural
        > kind terms. Thus, phenomenal concepts, for Chalmers, are unique in
        > that they are not constituted by any apriori functional role
        > specification, implying that the scientific explanation cannot get
        > off the ground, because there is nothing there for it to explain
        in
        > functional terms. This is his explanatory gap theory.
        >
        > Chalmers then uses this unique property of phenomenal
        > concepts in his argument against Materialism:
        >
        > (1) P&~Q is conceivable (cannot be ruled out apriori)
        > (2) If P&~Q is conceivable then P&~Q is 1-possible
        > (3) If P&~Q is 1-possible then either P&~Q is 2-possible or
        Russelian
        > Monism is true
        > (4) If P&~Q is 2-possible then Materialism is false
        > Therefore, either Materialism is false or Russelian Monism is true
        >
        > Q is a phenomenal truth, e.g., that someone is consciously
        > experiencing pain, while P is the conjunction of all microphysical
        > truths about the universe. P&~Q thus specifies a world identical
        to
        > ours in all microphysical respects, but which differs
        phenomenally.
        > We can see that premise one is substantive once we see that if we
        > replace Q with any non-phenomenal macroscopic truth, say, a truth
        > about water (=W), then P&~W will be inconceivable. That is, P will
        > tell us in topic neutral terms what plays the water role and all
        the
        > truths about the stuff that plays the water role. Thus, if we have
        > the concept of water, then we can infer apriori, from P, W. We
        could
        > thus rule P&~W out apriori. P&~Q cannot be ruled out apriori
        > precisely because the phenomenal term in Q has no apriori
        functional
        > role. Unfortunately I don't have time to explain this argument I
        just
        > present it show what's on the line, I will, however, be happy to
        > answer any questions concerning it.
        >
        > Levine's way out of the argument
        >
        > To escape the conclusion of the argument, Levine attempts to
        > trivialize the first premise so much so that nothing substantial
        can
        > be inferred from it. In particular, he wants to show that premise
        one
        > is true, not because of anything that is unique to phenomenal
        > concepts as Chalmers argues, but because practically nothing can
        be
        > ruled out apriori, due to the scarcity of apriori truths. In
        > particular, Levine wants to show that P&~W cannot be ruled out
        > apriori for the very same reason that P&~Q cannot be ruled out
        > apriori. Levine's conclusion then, is as follows: Clearly, one
        would
        > not conclude that there is an ontological gap between H20 and
        water
        > from the premise that P&~W can be ruled out apriori. So to one
        should
        > not conclude that there is an ontological gap between microphysics
        > and phenomenal properties, from the premise that P&~Q cannot be
        ruled
        > out apriori.
        > To clarify the disagreement over the extent of the set of apriori
        > truths, Levine introduces two concepts, `semantic form' and
        `logical
        > form'. For Levine (apart from some specific examples that I'll
        > mention later), the apriori truths are just the logical truths and
        so
        > he uses the term "logical form" to name the source of this apriori
        > knowledge. But despite the importance of this notion, he says
        nothing
        > about it, except that it is what allows us to know that logical
        > contradictions cannot express truths. To distinguish himself from
        > Chalmers, he notes that Chalmers believes that `semantic form' is
        > also a source of apriori knowledge. Levine is much clearer on
        > semantic form and defines it as:
        >
        > "…information concerning the application conditions of concepts.
        So
        > if I know that all bachelors are unmarried a priori it's because I
        > know that nothing counts as a bachelor unless its unmarried" (p41).
        >
        > So the semantic form for natural kind terms such as "water"
        consist
        > in their functional role specifications. With these concepts in
        hand,
        > Levine states the difference between himself and Chalmers, while
        > Chalmers believes that semantic form is a source of apriori
        > knowledge, Levine doesn't, Levine believes that the primary source
        is
        > logical form. If Levine is right, then even if P tells us that H20
        > plays the water role, we cannot then infer apriori, solely through
        > possessing the concept of water, that water is H20, for the water
        > role is not connected apriori to our term "Water". And so, if P&~Q
        is
        > conceivable in this sense, then nothing substantial can be
        concluded
        > from it. The question now becomes, why should we think that
        semantic
        > form is not a source of apriori knowledge?
        >
        > How does Levine Restrict apriority to just logical form?
        >
        > Levine's main tactic is to show that we have no reason to assume
        that
        > the class of apriori truths consists in anything more than the
        > logical truths. I will present what I take to be the two most
        > significant elements of Levine's attack. Firstly, his attempt to
        show
        > that the theoretical tool that Chalmers uses to uncover apriori
        > truths – conceptual analysis – is unfit for the task. And
        secondly,
        > his appeal to the causal theory of reference, which, in order to
        have
        > a coherent philosophy of language, is needed to supplement his
        > scepticism about the apriori.
        >
        > He believes that there is no reason to think that the results of
        > conceptual analysis are apriori
        >
        > When we do conceptual analysis, we give descriptions of all sorts
        of
        > possible worlds not involving a term T, ask ourselves what T
        refers
        > to or when T should be applied in these worlds, and then
        generalise
        > over our answers. Such generalizations are meant to capture
        general
        > facts about the application conditions, or meanings of T.
        Sentences
        > that describe such general facts are meant to be apriori, because
        > they survive the method of possible cases. That is, if a
        conceptual
        > analysis of our term "water" tells us that water is what plays the
        > water role, then we can test whether or not this sentence is
        apriori,
        > by trying stipulate a possible case in which we would consider it
        to
        > be false of water. If we cannot, then that is good evidence that
        the
        > truth of the sentence is not dependent on the way the world turns,
        so
        > it is not aposteriori, but it is dependent on how we wish to apply
        > the term in these worlds, which is something we know apriori.
        > Perhaps the most famous example, that is considered to be
        conceptual
        > analysis at least by defenders of conceptual analysis, is the
        > analysis of "water" given by Hilary Putnam . Putnam imagined that
        > somewhere in the galaxy there exists "twin-Earth" which is
        identical
        > to Earth, except that the transparent liquid in its seas rivers
        lakes
        > and rains (i.e. what plays its water role) is XYZ rather than H20.
        > Putnam then asked us if our word "water" would refer to the stuff
        > that plays their water role as well as what plays ours. Most
        people's
        > intuitions tell them that it doesn't, that the watery stuff on
        their
        > planet is not H20 and so is not water, it just seems like water.
        > Let's reformulate the example. Now imagine a possible world where
        we
        > find out that we were wrong in thinking that what plays our water
        > role is H20, as it is actually XYZ (so in a sense we find out that
        > twin-Earth is our earth). Would we say that "water" does not refer
        in
        > this possible world, or that it does, as it refers to XYZ? The
        > typical response is that it refers to XYZ.
        > What do these modal intuitions – intuitions about how to apply
        terms
        > in imagined cases – show us? They seem to show us what counts as
        > water. What counts as water is something like the transparent
        liquid
        > stuff that's in our seas rivers lakes and rains. Considering twin-
        > Earth as a planet in a different galaxy to ours, as Putnam invited
        us
        > to, brought out the indexical element – "our". And also
        that "water"
        > refers to the occupant of the water role and not the role property
        > itself (so that water is not identical to the water role but to
        what
        > plays the water role). And considering twin-Earth as a way our
        planet
        > might turn out to be shows us which associated properties – e.g.
        > transparency, liquidity, being in lakes and rivers etc. – go into
        > the `water role' specification that we use to pick out the water
        role
        > occupant.
        > Levine denies that such results represents apriori knowledge.
        > Concerning what the water role consists in, Levine believes that
        > these are merely our strongest beliefs about water, which are
        > empirical. And so, we use these empirical beliefs, which may well
        be
        > false of water, to pick out the extension of water in worlds.
        > Concerning the identity of water with the role occupant rather
        than
        > the role, Levine admits that empirical enquiry cannot help and
        that
        > only modal intuitions can inform us of this. But he argues that
        > whatever our concepts consist in, what constitutes them cannot be
        a
        > capacity for conceptual analysis, that is, this ability to reflect
        on
        > our use of concepts in possible worlds. This is because higher
        > animals have concepts, but cannot perform conceptual analysis.
        > Therefore, there is no reason to think that conceptual analysis
        tells
        > us about the requirements for the possession of concepts, be it
        > apriori knowledge or whatever. I'll come back to this.
        >
        > He appeals to a Causal theory of reference
        >
        > A theory of reference is one that answers the following question:
        Why
        > do our natural kind terms have the referents that they have?
        Chalmers
        > has an answer that invokes his broad conception of apriority:
        > Firstly, when speakers become competent with the use of the term,
        > they tacitly associate with it a functional role definition –
        water
        > is what plays the water role. Secondly, the world must play its
        part –
        > it must provide a referent that plays that functional role. Thus,
        > reference is determined by an apriori associated functional role
        and
        > the feature (if any) of the world that plays that role.
        Accordingly,
        > coming to know what a term refers to will consist in two parts –
        an
        > apriori and an aposteriori part. To know what "water" refers to,
        one
        > must become competent with the conventional use of "water", that
        is,
        > one must know that water is what plays the water role, and then,
        one
        > must empirically discover what actually plays the water role. The
        > former is apriori knowledge the latter aposteriori.
        > Levine needs to answer the question without appeal to apriority.
        This
        > is because, in taking the water role to be aposteriori, he must
        admit
        > that the water role specification might be false of water. In
        that
        > case, of course (possibly) false associated statements about water
        > cannot explain why "water" refers to water. So Levine instead
        appeals
        > to causal theories of reference.
        > Levine does not provide his own theory, all he says is that "What
        > determines extension […] are the external relations that obtain
        > between a representation and what it represents. On most theories
        the
        > relevant relation is something like causal covariation" (p54).
        > Nonetheless, we can illustrate what a theory of this form looks
        like
        > with Michael Devitt and Kim Sterelny's causal-historical theory.
        This
        > theory is developed out of the insights of Kripke and Putnam,
        which
        > is where these causal theories began. For D&S, a name or natural
        kind
        > term gets introduced in the presence of an object, or paradigmatic
        > instance or sample of some natural kind, which establishes a
        causal
        > link between the term and its extension. Those who introduce the
        term
        > go on to spread the term around the community, making more causal
        > chains, which all ultimately lead back to the dubbing ceremony and
        so
        > to the extension. What determines the reference of a name when I
        use
        > it is the causal chain that leads down to its referent. What
        > determines the reference of a natural kind term when I use it is
        the
        > causal chain which leads down to the paradigm instance(s) or sample
        > (s) of the kind which shares an underlying nature or sameness
        > relation to all other instances or samples of the term's
        extension.
        > These causal relations are external to the mind and are therefore
        > knowable only aposteriori. In other words, knowledge of such
        > reference determiners have not been drilled into the mind through
        the
        > practice of language acquisition at all such that we can tease out
        > info about them by conceptual analysis. Apriority thus plays no
        > significant role, for the causal theorist, in determining
        reference.
        >
        > My Critique of Levine
        >
        > Concerning the alternative theory of reference, I will argue that
        the
        > causal theory suffers from a problem that can only be resolved by
        > invoking a broad conception of apriority. Concerning conceptual
        > analysis, I will argue that Levine misconstrues it, and that once
        we
        > get it clear and use it to illuminate why logical truths might be
        > apriori as Levine assumes they are, we will see that Levine's
        > conception of the extent of apriority is arbitrary, while
        Chalmers'
        > isn't.
        >
        > Rejection of the Causal Theory (the qua-problem)
        >
        > Consider the dubbing ceremony for "water". What makes it the case
        > that the term refers to water qua natural kind as opposed to, say,
        > just the lake involved in the dubbing ceremony, so that "water" is
        > really a proper name? Or to illustrate the objection in another
        way;
        > what makes it the case that the term refers to water qua substance
        as
        > opposed to, say, the colour or weight of the sample water? So, on
        the
        > causal account, why doesn't "water" refer to an individual lake,
        or
        > to a particular colour or to a particular weight? Appeal to causal
        > links cannot solve this problem. It seems the only way out of the
        qua-
        > problem, is to leak some apriori knowledge in. So we know apriori
        > that water is a natural kind. But this isn't enough. What now
        > stops "water" from referring to, say, anything physical? Looks
        like
        > we will have to leak some more apriori knowledge in – perhaps that
        it
        > is a liquid. But this still won't do, for then why doesn't the
        term
        > refer to all kinds of liquid? It seems we need to say that it
        refers
        > to this kind of liquid while pointing at, and having in mind, say,
        > paradigmatic instances of our seas, rivers, lakes and rains.
        > What should we conclude from this about the explanatory worth of
        the
        > causal theory of reference? On the face of it, it seems that if we
        > keep hitting Levine with these qua-problems, we shall inevitably
        > arrive at Chalmers' theory of the apriori. By "water" we mean the
        > stuff that plays the water role, and so, we only need to reflect
        on
        > our concepts, without appeal to empirical information, to know
        that
        > water is what plays the water role. And once we are at this stage,
        it
        > seems as though we have enough to determine reference, for we have
        an
        > apriori water role specification, and then we just leave it up to
        the
        > world to provide a role occupant.
        >
        > Clarifying Conceptual Analysis and Drawing the apriori/aposteriori
        > divide
        >
        > Levine gives a particular interpretation to conceptual analysis.
        As
        > he says in a crucial footnote, Chalmers "is arguing from an
        analysis
        > of what it is to possess a concept to the claim that one must have
        a
        > priori knowledge of how extension is determined by context. It is
        > this inference that is being rejected" (p186). But what is
        > this "analysis of what it is to possess a concept" that Chalmers
        is
        > allegedly arguing from? Levine thinks that Chalmers is saying that
        > the ability to reflect on our practice with conceptual analysis is
        > what constitutes, or is at least necessary for, our linguistic
        > practice.
        > However, the ability to perform conceptual analysis does not
        > constitute our apriori knowledge, but rather, it makes explicit
        our
        > apriori knowledge. So what then, is the difference between we who
        are
        > capable of conceptual analysis, and other concept possessors, or
        more
        > specifically, other possessors of apriori knowledge, such as
        higher
        > animals? The difference is in the capacity to do conceptual
        analysis
        > of course, and this capacity involves more than the possession of
        > concepts – which is precisely what Levine has overlooked. To do
        > conceptual analysis, you must possess two other capacities:
        firstly,
        > a powerful imaginative capacity, and secondly, the ability to
        > describe complex imaginings with a complex linguistic system. Only
        > then can you describe complex possible worlds before asking
        yourself
        > what your concepts designate in those worlds. Without the latter
        two
        > capacities, the apriori knowledge that constitutes your concepts
        will
        > forever remain silent – trapped deep within the mind, as is the
        case
        > with higher animals.
        > With our methodology for uncovering apriori truths clear, we
        > should now turn to Levine's distinction between logical form and
        > semantic form. Levine expresses perplexity as to the nature of
        > logical form, that is, the source of our apriori knowledge of
        logical
        > truths, as he says:
        >
        > "I know a priori that I cannot both quit and not quit my job,
        because
        > I know that no statement of this form can express a truth; it's a
        > logical contradiction. How do I know that logical contradictions
        > cannot express truths? This is a deep question that I cannot take
        up
        > here" (p41).
        >
        > I believe that the source of our apriori knowledge of logical
        truths,
        > that is, logical form, simply consists in the semantic form of
        > logical terms. I am suggesting that we know that a logical
        > contradiction, say of the form "X and not X" cannot express a
        truth
        > due to our knowledge of the application conditions of logical
        terms.
        > Levine does not explain what he means by, "application
        > conditions", but read literally, the expression just means
        > information about the correct conditions under which we should
        apply
        > (or use) a term. And all terms, logical and non-logical alike,
        have
        > these. For all such terms, we have an intuitive understanding of
        the
        > conditions under which they ought to be used. So what might the
        > application conditions of logical terms be? One suggestion might
        be
        > that their truth tables capture their application conditions. And
        > there are two significant similarities between this and the
        > suggestion that "water"'s application conditions are captured by
        the
        > water role specification.
        > Firstly, while one must have a tacit understanding of the water
        role
        > to be able to use "water" with the community, similarly, it seems
        a
        > tacit understanding of the truth tables is necessary for one to be
        > able to apply logical terms correctly. Secondly, how did we come
        up
        > with truth tables? We did it the same way we came up with the
        water
        > role specification – conceptual analysis. Truth tables are
        idealized
        > generalizations over our actual use of logical connectives (in
        > worlds, if you like) that hold between propositions. This is why
        > propositions are replaced by variables within the tables. And most
        > importantly, the truth tables are correct if they survive the
        method
        > of possible cases. If you can imagine a situation that could be
        truly
        > described by a complex proposition whose truth is not predicted by
        > the truth tables of the relevant logical connectives, then those
        > truth tables will have been ruled out by conceptual analysis.
        > And so I can see no reason why we should be privileging logical
        terms
        > over non-logical terms when it comes to determining the extent of
        > apriori truths. Our methods for establishing that a logical truth
        is
        > apriori, is just conceptual analysis, and so there is no reason
        why
        > we cannot perform conceptual analysis on a sentence as a whole,
        > rather than just arbitrarily privileging its logical terms. Just
        to
        > drive the arbitrariness objection even further, Levine actually
        > thinks that the class of apriori truths is slightly bigger than
        the
        > set of logical truths. After his discussion of the modal intuition
        > that water is a role occupant rather than a role property, he says:
        >
        > "Finally it is open […] to admit that on the question of what kind
        of
        > thing water is—a role-property or a substance occupying the role—
        we
        > do have a priori access to the answer. But that isn't sufficient
        to
        > render [P&~W] conceptually impossible, which is the point at issue
        > here. So long as the specification of the role itself is not known
        a
        > priori, statement [P&~W] cannot be ruled out on a priori grounds"
        > (p62).
        >
        > But it is not a logical truth that water is an occupant property.
        So
        > then what else could make it true if not the fact that the
        > application conditions of "water" disallow us to be able to
        conceive
        > a world where water is a role property? So there seems to be no
        > substantial difference between this and the suggestion that the
        > application conditions of logical terms, as captured in their
        truth
        > tables, disallow us to conceive of a world where one can both
        quite
        > and not quit their job.
        > Levine also says:
        >
        > "Aristotle […] could not have failed to be human. Nothing that is
        not
        > human counts as the same individual as Aristotle […] it must be by
        > virtue of an a priori principle to the effect that "human being"
        is a
        > privileged sortal. That is, if the predicate applies at all to an
        > object, then it applies to that object in every possible world in
        > which the object exists" (p43).
        >
        > Once again, the sentence "if Aristotle is human then nothing
        that's
        > not human could be Aristotle", is not a logical truth. How then do
        we
        > know apriori that it is true? Surely by means of knowing the
        apriori
        > principle expressed the last sentence of the quote, which surely
        > captures an aspect of the application conditions which govern the
        use
        > of "human being" uncovered through conceptual analysis.
        > We can now pass this reasoning over to the crucial
        > sentence "water is what plays the water role". Is it really the
        case
        > that no empirical evidence could ever show us that water is not
        > actually the dominant clear, drinkable liquid in our seas, rivers,
        > lakes and rains? I can't see how. I can certainly imagine finding
        out
        > that water is not in the lakes, but that would just be to discover
        > that the dominant clear, drinkable liquid in our seas, rivers and
        > rains is not in our lakes. So the real apriori truth could be
        thought
        > of as:
        >
        > "Water is the stuff that exemplifies most of the basic properties
        > that I use to pick out its instances, such as being clear,
        drinkable,
        > and being the dominant stuff in our seas, rivers, lakes and rains".
        >
        > I see no reason to think that this is aposteriori, it seems clear
        to
        > me that, or at least something like it, exhausts our concept of
        > water.
        > And so, "water" does have a functional role specification, which
        can
        > be completely cashed in the above manner. If Chalmers is right
        that
        > this is required to give a reductive analysis of water, where
        science
        > tells us what more fundamental property plays this functional
        role,
        > then it seems we do have a significant difference between
        phenomenal
        > terms and ordinary natural kind terms. We can conceive of pain
        > obtaining independently of the instantiation of any functional
        role
        > property. Hence, no functional role specification can capture our
        > concept of pain. And if scientific reductions do require such
        > specifications, then we are left with an explanatory gap between
        > phenomenal properties and the physical world. I hope I have shown
        > that Levine's defense of Materialism has failed, and that if we
        wish
        > to escape the conceivability argument, then this is not the way.
        >
      • Chris Lofting
        Ron, ... See Hecht-Nielson s Confabulation Theory for arguments offering alternatives to the basic statistics models Hecht-Nielson (2007) Confabulation Theory
        Message 3 of 4 , Nov 8, 2007
        • 0 Attachment
          Ron,
          > -----Original Message-----
          > From: artificialintelligencegroup@yahoogroups.com
          > [mailto:artificialintelligencegroup@yahoogroups.com] On Behalf Of Ron
          > Sent: Thursday, 8 November 2007 11:23 AM
          > To: artificialintelligencegroup@yahoogroups.com
          > Subject: [Artificial Intelligence Group] Re: Apriori nature of mind/body
          > problem
          >
          > I was not sure if you wanted the machine to be conscious or if you
          > wanted it to be linked to people to help them to be conscious.
          >
          > It would seem to me that a generalized gaussian representation of
          > others is the father of an generalized gaussian representation of
          > self. Seeing your self then would be defined as a conscious
          > experience.
          >

          See Hecht-Nielson's Confabulation Theory for arguments offering alternatives
          to the basic statistics models

          Hecht-Nielson (2007)"Confabulation Theory : The Mechanism of Thought"
          Springer (2DVDs included to flesh out concepts)

          Or see http://scholarpedia.org/article/Confabulation_Theory

          He has not cottoned on to the self-referencing dynamic that encodes
          'meaning' categories but he is getting into the parallel processing context.
          He thus thinks is perspective is 'major' when the fact is it isn't if he
          read outside of his box! The only plus is his apparent development of
          software.

          His '4000+' thalmo-cortical modules reflect the IDM focus on
          self-referencing differentiate/integrate to give us a dimension of POSSIBLE
          meanings of the whole - different levels give us different number of modules
          moving from the general two of the dichotomy to 16+million after 24
          iterations of self-referencing where more elicits diminishing returns - the
          best windows being in 8 to 64 to 4096 categories (this reflects development
          along a hyperbolic form rather than exponential, and so N=N^2 as compared to
          2^N)

          As covered in the IDM focus, the full dimension get sorted into
          best-fit/worst-fit order by the local context. He also works off this in the
          form of ordering thalmo-cortical module responses into ordering of 'most
          excited' downwards.

          The IDM material focuses on the derived dimension representing classes of
          meaning and experiences are instances of such. As such each context, each
          experience, associates with a symbol mapping the class to the instance
          through a symbol. This comes out in Confab., theory as well but as a corpus
          of all possible symbols (words from many textual sources - IDM goes deeper
          but more generic in its focus on FEELINGS)

          Chris.
          http://members.iimetro.com.au/~lofting/myweb/introIDM.html
        • Kelvin Mcqueen
          ... I guess the former. ... The first sentence makes no sense to me. Best not to mention peoples names in an explanation. Concerning the second sentence, I
          Message 4 of 4 , Nov 9, 2007
          • 0 Attachment
            Quoting Ron <ronblue@...>:

            > I was not sure if you wanted the machine to be conscious or if you
            > wanted it to be linked to people to help them to be conscious.
            >

            I guess the former.

            > It would seem to me that a generalized gaussian representation of
            > others is the father of an generalized gaussian representation of
            > self. Seeing your self then would be defined as a conscious
            > experience.
            >

            The first sentence makes no sense to me. Best not to mention peoples names
            in an explanation. Concerning the second sentence, I don't think that that
            is a very good conceptual analysis of consciousness. Zombies can see
            themSELVES - they only need to look in the mirror.

            Cheers,
            Kelvin.

            >
            >
            > --- In artificialintelligencegroup@yahoogroups.com, "ilivetothink"
            > <mcqke464@...> wrote:
            > >
            > > Hi all,
            > >
            > > No idea if anyone here will be interested, but perhaps if you
            > think
            > > there is some possibilty of constructing a computer which could
            > > sustain consicousness, this will point out some difficulties...
            > >
            > > would be very interested in any
            > > questions/comments/criticisms/nitpicks...
            > >
            > > First section defines Materialism, second explains the apriori
            > nature
            > > of the explanatory gap, third is Levine's critique of apriority,
            > > fourth is my critique of Levine.
            > >
            > > Cheers,
            > > Kelvin
            > >
            > >
            > >
            > >
            > >
            > >
            > > Introduction
            > >
            > > Joseph Levine is a Materialist. In other words, he believes
            > > that "Only non-mental properties are instantiated in a basic way;
            > all
            > > mental properties are instantiated by being realized by the
            > > instantiation of other, non-mental properties" (p21). By "basic"
            > > Levine means "irreducible", and so the basic properties (and
            > > entities) are the ones that will be mentioned in a completed
            > > fundamental physics. `Realization', for Levine, is a relation
            > between
            > > properties, distinct from the relations of identity and of lawful
            > > correlation. He defines it as follows: "The instantiation of
            > property
            > > is A is realized by the instantiation of property B just in case
            > the
            > > very fact alone of B's instantiation constitutes the instantiation
            > of
            > > A [but not vice versa]" (p13). I have added "but not vice versa"
            > into
            > > Levine's definition to highlight the crucial distinction between
            > > realization and identity. Levine's favourite examples of
            > realization
            > > come from computer science, where functional properties, such as
            > > computer programs, are physically realized by the electronic
            > > mechanisms in the computer. Consider a chess program. The reason
            > the
            > > relation between the program and the electronic mechanisms that
            > > sustain it is not one of identity is because a chess program could
            > be
            > > realised by something else, such as a person moving around chess
            > > pieces on a chessboard. The same holds for mental properties such
            > as
            > > intelligence or pain. We might say that the firing of C-fibers
            > > realizes pain in human beings, but in other creatures that feel
            > pain,
            > > it may be something quite different.
            > > If Materialism is true then we should be able to explain how
            > > mental properties are realized by non-mental properties, or in
            > other
            > > words, how mental properties reduce to non-mental properties.
            > Levine
            > > believes we are well on our way in doing this for rationality and
            > > intentionality, but he thinks there is a major problem for
            > conscious
            > > experience, or more technically, phenomenal properties. As defined
            > by
            > > Thomas Nagel, a property is phenomenal just in case there is
            > > something it is like to have that property. So pain is a
            > phenomenal
            > > property because there is something it is like to feel pain.
            > Levine
            > > thinks that the greatest challenge for Materialism is the so-
            > > called "conceivability argument", particularly David Chalmers'
            > modern
            > > defense of it. This argument infers an ontological gap between
            > > consciousness and non-mental physical properties, from an
            > explanatory
            > > gap, that is, from a theory explaining why we cannot reductively
            > > explain consciousness in non-mental physical terms. While Levine
            > is
            > > very sympathetic with the explanatory gap (he actually invented
            > the
            > > notion 24 years ago), he refuses to accept an ontological gap.
            > > Chalmers' explanatory gap theory is set up in a way that makes it
            > > easy for him to infer an ontological gap. Thus, Levine criticizes
            > the
            > > semantic assumptions involved in Chalmers' theory and constructs a
            > > quite different semantics of his own. In this paper, I will argue
            > > that Levine's philosophy of language cannot work and so neither
            > can
            > > his arguments against Chalmers' from which I will conclude that
            > > Levine has failed to save Materialism from the threat of the
            > > conceivability argument.
            > >
            > > The object of Levine's attack: Chalmers' explanatory gap theory
            > >
            > > Let's begin by looking at Chalmers' explanatory gap theory. As
            > > mentioned, Materialism requires reductive analyses, because it
            > does
            > > not want to treat phenomenal properties as basic – they must be
            > > reducible to their physical realizers. But for Chalmers, reductive
            > > explanations involve two specific requirements, which reductions
            > of
            > > phenomenal properties can't seem to meet, that is, conceptual
            > > analysis and scientific explanation. I shall illustrate with the
            > > simple example of the reduction of water to H20:
            > >
            > > [Conceptual Analysis:] Water is what plays the water role (Apriori)
            > > [Scientific Explanation:] H20 is what plays the water role
            > > (Aposteriori)
            > > [Reductive Conclusion:] Water is (identical to) H20.
            > >
            > > The first premise consists in cashing out our concept of water,
            > that
            > > is, what we mean when we speak of water, so that we are clear on
            > what
            > > we want to reductively explain. Chalmers takes natural kind
            > concepts
            > > like "water" to be functional or causal role concepts, and so, a
            > > complete conceptual analysis will tell us what the water role is,
            > and
            > > that water is what plays that role. A rough preliminary conceptual
            > > analysis might tell us that water is the dominant clear, drinkable
            > > liquid in our seas, rivers, lakes and rains. Given that conceptual
            > > analysis tells us about what we already know tacitly in virtue of
            > > linguistic competence, this truth about water is known apriori.
            > > Levine disagrees sharply with this, and so I shall come back to
            > > conceptual analysis in detail later.
            > > Concerning the second premise, this is where scientists tell us
            > that
            > > H20 is the dominant clear, drinkable liquid in our seas, rivers,
            > > lakes and rains. As we shall soon see, for Chalmers' argument
            > against
            > > Materialism to work, Chalmers needs it to be the case that we can
            > (at
            > > least in principle) be told that H20 plays this functional role
            > > purely in microphysical terms. So for example, a complete
            > > microphysical theory of the universe will tell us the geometrical
            > > properties of H20 molecules in space-time, from which, given
            > > possession of the concept of liquidity, we should be able to infer
            > > that H20 molecules exemplify liquidity. It will also tell us about
            > > spatial locations of H20, from which we should be able to infer,
            > > given possession of the concepts of lakes and rains that H20
            > > molecules are in our lakes and fall from our skies. We thus get
            > told
            > > in terms neutral to macrophysical and microphysical vocabulary
            > (e.g.
            > > in mathematical, geometrical and causal terms), that H20 plays the
            > > water role, from which we can infer, given our apriori premise,
            > that
            > > water is H20.
            > > For Chalmers, unless we can do this for phenomenal properties,
            > then
            > > we are not going to get a reduction, and we will have an
            > explanatory
            > > gap. In that case, Materialism may still be true, we just won't be
            > > able to establish it. So Materialism needs something like:
            > >
            > > [Conceptual Analysis:] Pain is what plays the pain role
            > (Apriori???)
            > > [Scientific Explanation:] C-Fibers firing is what (at least in
            > > humans) plays the pain role (Aposteriori)
            > > [Reductive Conclusion:] Pain is (realised in humans by) the firing
            > of
            > > C-fibers
            > >
            > > Perhaps premise one tells us that pain is what typically causes
            > > certain behaviours like writhing while being caused by certain
            > inputs
            > > such as tissue damage. The problem here, for Chalmers, is that
            > this
            > > truth about pain is not apriori. Chalmers backs this up by saying
            > > that whatever constitutes our concept of pain, it is constituted
            > in
            > > such a way that we can conceive of the instantiation of any such
            > > functions in the absence of pain. For example we can conceive of
            > > creatures who constantly instantiate this so-called pain role but
            > > never feel pain. The same cannot be said for non-phenomenal
            > natural
            > > kind terms. Thus, phenomenal concepts, for Chalmers, are unique in
            > > that they are not constituted by any apriori functional role
            > > specification, implying that the scientific explanation cannot get
            > > off the ground, because there is nothing there for it to explain
            > in
            > > functional terms. This is his explanatory gap theory.
            > >
            > > Chalmers then uses this unique property of phenomenal
            > > concepts in his argument against Materialism:
            > >
            > > (1) P&~Q is conceivable (cannot be ruled out apriori)
            > > (2) If P&~Q is conceivable then P&~Q is 1-possible
            > > (3) If P&~Q is 1-possible then either P&~Q is 2-possible or
            > Russelian
            > > Monism is true
            > > (4) If P&~Q is 2-possible then Materialism is false
            > > Therefore, either Materialism is false or Russelian Monism is true
            > >
            > > Q is a phenomenal truth, e.g., that someone is consciously
            > > experiencing pain, while P is the conjunction of all microphysical
            > > truths about the universe. P&~Q thus specifies a world identical
            > to
            > > ours in all microphysical respects, but which differs
            > phenomenally.
            > > We can see that premise one is substantive once we see that if we
            > > replace Q with any non-phenomenal macroscopic truth, say, a truth
            > > about water (=W), then P&~W will be inconceivable. That is, P will
            > > tell us in topic neutral terms what plays the water role and all
            > the
            > > truths about the stuff that plays the water role. Thus, if we have
            > > the concept of water, then we can infer apriori, from P, W. We
            > could
            > > thus rule P&~W out apriori. P&~Q cannot be ruled out apriori
            > > precisely because the phenomenal term in Q has no apriori
            > functional
            > > role. Unfortunately I don't have time to explain this argument I
            > just
            > > present it show what's on the line, I will, however, be happy to
            > > answer any questions concerning it.
            > >
            > > Levine's way out of the argument
            > >
            > > To escape the conclusion of the argument, Levine attempts to
            > > trivialize the first premise so much so that nothing substantial
            > can
            > > be inferred from it. In particular, he wants to show that premise
            > one
            > > is true, not because of anything that is unique to phenomenal
            > > concepts as Chalmers argues, but because practically nothing can
            > be
            > > ruled out apriori, due to the scarcity of apriori truths. In
            > > particular, Levine wants to show that P&~W cannot be ruled out
            > > apriori for the very same reason that P&~Q cannot be ruled out
            > > apriori. Levine's conclusion then, is as follows: Clearly, one
            > would
            > > not conclude that there is an ontological gap between H20 and
            > water
            > > from the premise that P&~W can be ruled out apriori. So to one
            > should
            > > not conclude that there is an ontological gap between microphysics
            > > and phenomenal properties, from the premise that P&~Q cannot be
            > ruled
            > > out apriori.
            > > To clarify the disagreement over the extent of the set of apriori
            > > truths, Levine introduces two concepts, `semantic form' and
            > `logical
            > > form'. For Levine (apart from some specific examples that I'll
            > > mention later), the apriori truths are just the logical truths and
            > so
            > > he uses the term "logical form" to name the source of this apriori
            > > knowledge. But despite the importance of this notion, he says
            > nothing
            > > about it, except that it is what allows us to know that logical
            > > contradictions cannot express truths. To distinguish himself from
            > > Chalmers, he notes that Chalmers believes that `semantic form' is
            > > also a source of apriori knowledge. Levine is much clearer on
            > > semantic form and defines it as:
            > >
            > > "…information concerning the application conditions of concepts.
            > So
            > > if I know that all bachelors are unmarried a priori it's because I
            > > know that nothing counts as a bachelor unless its unmarried" (p41).
            > >
            > > So the semantic form for natural kind terms such as "water"
            > consist
            > > in their functional role specifications. With these concepts in
            > hand,
            > > Levine states the difference between himself and Chalmers, while
            > > Chalmers believes that semantic form is a source of apriori
            > > knowledge, Levine doesn't, Levine believes that the primary source
            > is
            > > logical form. If Levine is right, then even if P tells us that H20
            > > plays the water role, we cannot then infer apriori, solely through
            > > possessing the concept of water, that water is H20, for the water
            > > role is not connected apriori to our term "Water". And so, if P&~Q
            > is
            > > conceivable in this sense, then nothing substantial can be
            > concluded
            > > from it. The question now becomes, why should we think that
            > semantic
            > > form is not a source of apriori knowledge?
            > >
            > > How does Levine Restrict apriority to just logical form?
            > >
            > > Levine's main tactic is to show that we have no reason to assume
            > that
            > > the class of apriori truths consists in anything more than the
            > > logical truths. I will present what I take to be the two most
            > > significant elements of Levine's attack. Firstly, his attempt to
            > show
            > > that the theoretical tool that Chalmers uses to uncover apriori
            > > truths – conceptual analysis – is unfit for the task. And
            > secondly,
            > > his appeal to the causal theory of reference, which, in order to
            > have
            > > a coherent philosophy of language, is needed to supplement his
            > > scepticism about the apriori.
            > >
            > > He believes that there is no reason to think that the results of
            > > conceptual analysis are apriori
            > >
            > > When we do conceptual analysis, we give descriptions of all sorts
            > of
            > > possible worlds not involving a term T, ask ourselves what T
            > refers
            > > to or when T should be applied in these worlds, and then
            > generalise
            > > over our answers. Such generalizations are meant to capture
            > general
            > > facts about the application conditions, or meanings of T.
            > Sentences
            > > that describe such general facts are meant to be apriori, because
            > > they survive the method of possible cases. That is, if a
            > conceptual
            > > analysis of our term "water" tells us that water is what plays the
            > > water role, then we can test whether or not this sentence is
            > apriori,
            > > by trying stipulate a possible case in which we would consider it
            > to
            > > be false of water. If we cannot, then that is good evidence that
            > the
            > > truth of the sentence is not dependent on the way the world turns,
            > so
            > > it is not aposteriori, but it is dependent on how we wish to apply
            > > the term in these worlds, which is something we know apriori.
            > > Perhaps the most famous example, that is considered to be
            > conceptual
            > > analysis at least by defenders of conceptual analysis, is the
            > > analysis of "water" given by Hilary Putnam . Putnam imagined that
            > > somewhere in the galaxy there exists "twin-Earth" which is
            > identical
            > > to Earth, except that the transparent liquid in its seas rivers
            > lakes
            > > and rains (i.e. what plays its water role) is XYZ rather than H20.
            > > Putnam then asked us if our word "water" would refer to the stuff
            > > that plays their water role as well as what plays ours. Most
            > people's
            > > intuitions tell them that it doesn't, that the watery stuff on
            > their
            > > planet is not H20 and so is not water, it just seems like water.
            > > Let's reformulate the example. Now imagine a possible world where
            > we
            > > find out that we were wrong in thinking that what plays our water
            > > role is H20, as it is actually XYZ (so in a sense we find out that
            > > twin-Earth is our earth). Would we say that "water" does not refer
            > in
            > > this possible world, or that it does, as it refers to XYZ? The
            > > typical response is that it refers to XYZ.
            > > What do these modal intuitions – intuitions about how to apply
            > terms
            > > in imagined cases – show us? They seem to show us what counts as
            > > water. What counts as water is something like the transparent
            > liquid
            > > stuff that's in our seas rivers lakes and rains. Considering twin-
            > > Earth as a planet in a different galaxy to ours, as Putnam invited
            > us
            > > to, brought out the indexical element – "our". And also
            > that "water"
            > > refers to the occupant of the water role and not the role property
            > > itself (so that water is not identical to the water role but to
            > what
            > > plays the water role). And considering twin-Earth as a way our
            > planet
            > > might turn out to be shows us which associated properties – e.g.
            > > transparency, liquidity, being in lakes and rivers etc. – go into
            > > the `water role' specification that we use to pick out the water
            > role
            > > occupant.
            > > Levine denies that such results represents apriori knowledge.
            > > Concerning what the water role consists in, Levine believes that
            > > these are merely our strongest beliefs about water, which are
            > > empirical. And so, we use these empirical beliefs, which may well
            > be
            > > false of water, to pick out the extension of water in worlds.
            > > Concerning the identity of water with the role occupant rather
            > than
            > > the role, Levine admits that empirical enquiry cannot help and
            > that
            > > only modal intuitions can inform us of this. But he argues that
            > > whatever our concepts consist in, what constitutes them cannot be
            > a
            > > capacity for conceptual analysis, that is, this ability to reflect
            > on
            > > our use of concepts in possible worlds. This is because higher
            > > animals have concepts, but cannot perform conceptual analysis.
            > > Therefore, there is no reason to think that conceptual analysis
            > tells
            > > us about the requirements for the possession of concepts, be it
            > > apriori knowledge or whatever. I'll come back to this.
            > >
            > > He appeals to a Causal theory of reference
            > >
            > > A theory of reference is one that answers the following question:
            > Why
            > > do our natural kind terms have the referents that they have?
            > Chalmers
            > > has an answer that invokes his broad conception of apriority:
            > > Firstly, when speakers become competent with the use of the term,
            > > they tacitly associate with it a functional role definition –
            > water
            > > is what plays the water role. Secondly, the world must play its
            > part –
            > > it must provide a referent that plays that functional role. Thus,
            > > reference is determined by an apriori associated functional role
            > and
            > > the feature (if any) of the world that plays that role.
            > Accordingly,
            > > coming to know what a term refers to will consist in two parts –
            > an
            > > apriori and an aposteriori part. To know what "water" refers to,
            > one
            > > must become competent with the conventional use of "water", that
            > is,
            > > one must know that water is what plays the water role, and then,
            > one
            > > must empirically discover what actually plays the water role. The
            > > former is apriori knowledge the latter aposteriori.
            > > Levine needs to answer the question without appeal to apriority.
            > This
            > > is because, in taking the water role to be aposteriori, he must
            > admit
            > > that the water role specification might be false of water. In
            > that
            > > case, of course (possibly) false associated statements about water
            > > cannot explain why "water" refers to water. So Levine instead
            > appeals
            > > to causal theories of reference.
            > > Levine does not provide his own theory, all he says is that "What
            > > determines extension […] are the external relations that obtain
            > > between a representation and what it represents. On most theories
            > the
            > > relevant relation is something like causal covariation" (p54).
            > > Nonetheless, we can illustrate what a theory of this form looks
            > like
            > > with Michael Devitt and Kim Sterelny's causal-historical theory.
            > This
            > > theory is developed out of the insights of Kripke and Putnam,
            > which
            > > is where these causal theories began. For D&S, a name or natural
            > kind
            > > term gets introduced in the presence of an object, or paradigmatic
            > > instance or sample of some natural kind, which establishes a
            > causal
            > > link between the term and its extension. Those who introduce the
            > term
            > > go on to spread the term around the community, making more causal
            > > chains, which all ultimately lead back to the dubbing ceremony and
            > so
            > > to the extension. What determines the reference of a name when I
            > use
            > > it is the causal chain that leads down to its referent. What
            > > determines the reference of a natural kind term when I use it is
            > the
            > > causal chain which leads down to the paradigm instance(s) or sample
            > > (s) of the kind which shares an underlying nature or sameness
            > > relation to all other instances or samples of the term's
            > extension.
            > > These causal relations are external to the mind and are therefore
            > > knowable only aposteriori. In other words, knowledge of such
            > > reference determiners have not been drilled into the mind through
            > the
            > > practice of language acquisition at all such that we can tease out
            > > info about them by conceptual analysis. Apriority thus plays no
            > > significant role, for the causal theorist, in determining
            > reference.
            > >
            > > My Critique of Levine
            > >
            > > Concerning the alternative theory of reference, I will argue that
            > the
            > > causal theory suffers from a problem that can only be resolved by
            > > invoking a broad conception of apriority. Concerning conceptual
            > > analysis, I will argue that Levine misconstrues it, and that once
            > we
            > > get it clear and use it to illuminate why logical truths might be
            > > apriori as Levine assumes they are, we will see that Levine's
            > > conception of the extent of apriority is arbitrary, while
            > Chalmers'
            > > isn't.
            > >
            > > Rejection of the Causal Theory (the qua-problem)
            > >
            > > Consider the dubbing ceremony for "water". What makes it the case
            > > that the term refers to water qua natural kind as opposed to, say,
            > > just the lake involved in the dubbing ceremony, so that "water" is
            > > really a proper name? Or to illustrate the objection in another
            > way;
            > > what makes it the case that the term refers to water qua substance
            > as
            > > opposed to, say, the colour or weight of the sample water? So, on
            > the
            > > causal account, why doesn't "water" refer to an individual lake,
            > or
            > > to a particular colour or to a particular weight? Appeal to causal
            > > links cannot solve this problem. It seems the only way out of the
            > qua-
            > > problem, is to leak some apriori knowledge in. So we know apriori
            > > that water is a natural kind. But this isn't enough. What now
            > > stops "water" from referring to, say, anything physical? Looks
            > like
            > > we will have to leak some more apriori knowledge in – perhaps that
            > it
            > > is a liquid. But this still won't do, for then why doesn't the
            > term
            > > refer to all kinds of liquid? It seems we need to say that it
            > refers
            > > to this kind of liquid while pointing at, and having in mind, say,
            > > paradigmatic instances of our seas, rivers, lakes and rains.
            > > What should we conclude from this about the explanatory worth of
            > the
            > > causal theory of reference? On the face of it, it seems that if we
            > > keep hitting Levine with these qua-problems, we shall inevitably
            > > arrive at Chalmers' theory of the apriori. By "water" we mean the
            > > stuff that plays the water role, and so, we only need to reflect
            > on
            > > our concepts, without appeal to empirical information, to know
            > that
            > > water is what plays the water role. And once we are at this stage,
            > it
            > > seems as though we have enough to determine reference, for we have
            > an
            > > apriori water role specification, and then we just leave it up to
            > the
            > > world to provide a role occupant.
            > >
            > > Clarifying Conceptual Analysis and Drawing the apriori/aposteriori
            > > divide
            > >
            > > Levine gives a particular interpretation to conceptual analysis.
            > As
            > > he says in a crucial footnote, Chalmers "is arguing from an
            > analysis
            > > of what it is to possess a concept to the claim that one must have
            > a
            > > priori knowledge of how extension is determined by context. It is
            > > this inference that is being rejected" (p186). But what is
            > > this "analysis of what it is to possess a concept" that Chalmers
            > is
            > > allegedly arguing from? Levine thinks that Chalmers is saying that
            > > the ability to reflect on our practice with conceptual analysis is
            > > what constitutes, or is at least necessary for, our linguistic
            > > practice.
            > > However, the ability to perform conceptual analysis does not
            > > constitute our apriori knowledge, but rather, it makes explicit
            > our
            > > apriori knowledge. So what then, is the difference between we who
            > are
            > > capable of conceptual analysis, and other concept possessors, or
            > more
            > > specifically, other possessors of apriori knowledge, such as
            > higher
            > > animals? The difference is in the capacity to do conceptual
            > analysis
            > > of course, and this capacity involves more than the possession of
            > > concepts – which is precisely what Levine has overlooked. To do
            > > conceptual analysis, you must possess two other capacities:
            > firstly,
            > > a powerful imaginative capacity, and secondly, the ability to
            > > describe complex imaginings with a complex linguistic system. Only
            > > then can you describe complex possible worlds before asking
            > yourself
            > > what your concepts designate in those worlds. Without the latter
            > two
            > > capacities, the apriori knowledge that constitutes your concepts
            > will
            > > forever remain silent – trapped deep within the mind, as is the
            > case
            > > with higher animals.
            > > With our methodology for uncovering apriori truths clear, we
            > > should now turn to Levine's distinction between logical form and
            > > semantic form. Levine expresses perplexity as to the nature of
            > > logical form, that is, the source of our apriori knowledge of
            > logical
            > > truths, as he says:
            > >
            > > "I know a priori that I cannot both quit and not quit my job,
            > because
            > > I know that no statement of this form can express a truth; it's a
            > > logical contradiction. How do I know that logical contradictions
            > > cannot express truths? This is a deep question that I cannot take
            > up
            > > here" (p41).
            > >
            > > I believe that the source of our apriori knowledge of logical
            > truths,
            > > that is, logical form, simply consists in the semantic form of
            > > logical terms. I am suggesting that we know that a logical
            > > contradiction, say of the form "X and not X" cannot express a
            > truth
            > > due to our knowledge of the application conditions of logical
            > terms.
            > > Levine does not explain what he means by, "application
            > > conditions", but read literally, the expression just means
            > > information about the correct conditions under which we should
            > apply
            > > (or use) a term. And all terms, logical and non-logical alike,
            > have
            > > these. For all such terms, we have an intuitive understanding of
            > the
            > > conditions under which they ought to be used. So what might the
            > > application conditions of logical terms be? One suggestion might
            > be
            > > that their truth tables capture their application conditions. And
            > > there are two significant similarities between this and the
            > > suggestion that "water"'s application conditions are captured by
            > the
            > > water role specification.
            > > Firstly, while one must have a tacit understanding of the water
            > role
            > > to be able to use "water" with the community, similarly, it seems
            > a
            > > tacit understanding of the truth tables is necessary for one to be
            > > able to apply logical terms correctly. Secondly, how did we come
            > up
            > > with truth tables? We did it the same way we came up with the
            > water
            > > role specification – conceptual analysis. Truth tables are
            > idealized
            > > generalizations over our actual use of logical connectives (in
            > > worlds, if you like) that hold between propositions. This is why
            > > propositions are replaced by variables within the tables. And most
            > > importantly, the truth tables are correct if they survive the
            > method
            > > of possible cases. If you can imagine a situation that could be
            > truly
            > > described by a complex proposition whose truth is not predicted by
            > > the truth tables of the relevant logical connectives, then those
            > > truth tables will have been ruled out by conceptual analysis.
            > > And so I can see no reason why we should be privileging logical
            > terms
            > > over non-logical terms when it comes to determining the extent of
            > > apriori truths. Our methods for establishing that a logical truth
            > is
            > > apriori, is just conceptual analysis, and so there is no reason
            > why
            > > we cannot perform conceptual analysis on a sentence as a whole,
            > > rather than just arbitrarily privileging its logical terms. Just
            > to
            > > drive the arbitrariness objection even further, Levine actually
            > > thinks that the class of apriori truths is slightly bigger than
            > the
            > > set of logical truths. After his discussion of the modal intuition
            > > that water is a role occupant rather than a role property, he says:
            > >
            > > "Finally it is open […] to admit that on the question of what kind
            > of
            > > thing water is—a role-property or a substance occupying the role—
            > we
            > > do have a priori access to the answer. But that isn't sufficient
            > to
            > > render [P&~W] conceptually impossible, which is the point at issue
            > > here. So long as the specification of the role itself is not known
            > a
            > > priori, statement [P&~W] cannot be ruled out on a priori grounds"
            > > (p62).
            > >
            > > But it is not a logical truth that water is an occupant property.
            > So
            > > then what else could make it true if not the fact that the
            > > application conditions of "water" disallow us to be able to
            > conceive
            > > a world where water is a role property? So there seems to be no
            > > substantial difference between this and the suggestion that the
            > > application conditions of logical terms, as captured in their
            > truth
            > > tables, disallow us to conceive of a world where one can both
            > quite
            > > and not quit their job.
            > > Levine also says:
            > >
            > > "Aristotle […] could not have failed to be human. Nothing that is
            > not
            > > human counts as the same individual as Aristotle […] it must be by
            > > virtue of an a priori principle to the effect that "human being"
            > is a
            > > privileged sortal. That is, if the predicate applies at all to an
            > > object, then it applies to that object in every possible world in
            > > which the object exists" (p43).
            > >
            > > Once again, the sentence "if Aristotle is human then nothing
            > that's
            > > not human could be Aristotle", is not a logical truth. How then do
            > we
            > > know apriori that it is true? Surely by means of knowing the
            > apriori
            > > principle expressed the last sentence of the quote, which surely
            > > captures an aspect of the application conditions which govern the
            > use
            > > of "human being" uncovered through conceptual analysis.
            > > We can now pass this reasoning over to the crucial
            > > sentence "water is what plays the water role". Is it really the
            > case
            > > that no empirical evidence could ever show us that water is not
            > > actually the dominant clear, drinkable liquid in our seas, rivers,
            > > lakes and rains? I can't see how. I can certainly imagine finding
            > out
            > > that water is not in the lakes, but that would just be to discover
            > > that the dominant clear, drinkable liquid in our seas, rivers and
            > > rains is not in our lakes. So the real apriori truth could be
            > thought
            > > of as:
            > >
            > > "Water is the stuff that exemplifies most of the basic properties
            > > that I use to pick out its instances, such as being clear,
            > drinkable,
            > > and being the dominant stuff in our seas, rivers, lakes and rains".
            > >
            > > I see no reason to think that this is aposteriori, it seems clear
            > to
            > > me that, or at least something like it, exhausts our concept of
            > > water.
            > > And so, "water" does have a functional role specification, which
            > can
            > > be completely cashed in the above manner. If Chalmers is right
            > that
            > > this is required to give a reductive analysis of water, where
            > science
            > > tells us what more fundamental property plays this functional
            > role,
            > > then it seems we do have a significant difference between
            > phenomenal
            > > terms and ordinary natural kind terms. We can conceive of pain
            > > obtaining independently of the instantiation of any functional
            > role
            > > property. Hence, no functional role specification can capture our
            > > concept of pain. And if scientific reductions do require such
            > > specifications, then we are left with an explanatory gap between
            > > phenomenal properties and the physical world. I hope I have shown
            > > that Levine's defense of Materialism has failed, and that if we
            > wish
            > > to escape the conceivability argument, then this is not the way.
            > >
            >
            >
            >
          Your message has been successfully submitted and would be delivered to recipients shortly.