Reduction, Emergence, Etc.


In Ken Aizawa’s simple (and in my opinion compelling) argument against the extended mind hypothesis applied to consciousness, he assumed that properties and relations of lower-level enties determine those of higher-level entities.  In the comments, Adam Arico asked about emergence and Flora Carpenter further elaborated as follows:

      Higher level states may have emergent features by virtue of their functional organization 
      that are not determined among lower level parts, and this higher level organization 
      (e.g. consciousness) uniquely influences the system.

Adam and Flora’s question brings out some important themes in contemporary metaphysics of mind.  I thought I’d add the following points to Ken’s replies:

1. As Ken pointed out, the question of emergence is orthogonal to the conclusion of Ken’s argument.  Even if consciousness is emergent in the way described by Flora, it doesn’t follow that consciousness “extends beyond the brain”, as Alva Noe argued at the last Central APA.

2. The big debate about the relationship between higher and lower levels is usually between reductionism (e.g., Armstrong, Kim, David Lewis) and anti-reductionism (e.g., Fodor, Block).  In this debate, “emergence” is usually employed to insult opponents (as in, “you anti-reductionists are nothing but a bunch of emergentists!” “No, we ain’t!”).

3. Emergentism was semi-popular in the early 20th century.  But that was before the demise of vitalism.  These days, few metaphysicians would want to get caught advocating the kind of emergence described by Flora.  The reasons are quite simple:  (a) vitalism is pretty dead; (b) where do the emergent features come from if not from the features of the constituents of the whole plus the way the constituents are organized?  Speaking of “downward causation” from the whole’s properties to its parts’ properties sounds “spooky” and “magical” (I take terms in scare quotes directly from the literature), and no one seems to know how to tell a compelling philosophical story about it.

4. An interesting alternative to the traditional reductionism/anti-reductionism debate is the view held by C.B. Martin and John Heil (cf. Heil’s book, From an Ontological Point of View, OUP, 2003).  According to them, there is no issue of reduction vs. anti-reduction vs. (God forbid) emergence because there aren’t any levels to reduce or emerge.  There are levels of predicates and explanations, but not levels of properties.  All properties and relations are fundamental; there are no higher level ones.  And without higher level properties, the question of whether they reduce to or emerge from the lower level ones doesn’t arise.  Interestingly, Martin argues that there is still such a thing as emergence, but it’s something that happens at the fundamental level, that is, the properties of the components may change once they are organized together in special ways.  See his great paper, Martin, C. B. ‘The Need for Properties: The Road to Pythagoreanism and Back’, Synthese 112 (1997): 193-231.

0 Comments

  1. Adam Arico

    Let me just go on record as saying that my question on emergence was not intended to be insultory, only clarificatory. I wasn’t implying any sort of mystical vitalism in Ken’s brief explanation, nor was I aware that “emergence” has become such a dirty word in contemporary metaphysics of mind. Thanks, Gualtiero, for pointing out the error of my ways.

  2. gualtiero piccinini

    I didn’t mean to suggest you, Adam, are a vitalist or made any big mistake. BTW, there are still people who defend notions of emergence as useful versions of non-reductive materialism. As usual in philosophy, whether a notion is plausible or not depends on how it is understood (what its committments are). Emergence need not be spooky (cf. the emergence that C.B. Martin talks about), though it is often assumed to be.

  3. I thought I should mention following possibility which comes from logic.
    Namely we know that “A is B” doesn’t necessarily imply “A is merely B”.
    So even if we all agree that “whole is dynamics of parts”, it doesn’t have to mean “whole is merely the dynamics of its parts”, or “whole is nothing but the dynamics of its parts”.

    Reductionism is too optimistic in my opinion. It takes “A is B”, to mean “A is merely B”.

  4. Carl Gillett

    Gaultiero, just a quick comment, since I dropped into the site but am way busy with writing.

    I am a little lost why you don’t think Heil and Martin are ontological reductionists? They argue we should only belive in one layer, in other words that everything reduces, ontologically, to this one layer.

    Why then are they not a species of reductionists, rather than some further alternative? Best Carl

  5. Eric Thomson

    A couple of points. Emergence has a perfectly vanilla use that describes features of systems that emerge due to interactions of parts of the system, features would not obtain if the parts were not interacting (this is typically illustrated using Conway’s game of life). The point is that interesting dynamics and forms emerge from the interaction of simple things using simple rules. Since the brain consists of complex things interacting via complex rules, that makes our jobs as neuroscientists possible (and hard). Just as it is helpful to describe certain scenarios in the Game of Life using ‘higher level’ descriptions of ’emergent’ features (e.g., swimmers moving along etc), the same will likely happen in brains.

    The issue of possible top-down effects might be more complicated than commentators are allowing. I think Bechtel uses an example of explaining the change in position of a molecule by appeal to ethology (e.g., the chlorophyl molecule at the tip of the venus fly trap moved because a fly landed on the appropriate trigger in the plant). I realize this example is uninteresting, but might we be able to pump it for ideas?

    As for the claim that there aren’t any levels, that seems quite odd. How do they account for the explanatory assymmetry between, say thermodynamics and statistical mechanics. That is, one can be (roughly) derived from the other, but not vice versa.

  6. Eric Thomson

    I asked how the anti-levels people would account for explanatory assymmetry: the answer is in your original summary. You say they admit of explanatory/predicate levels, but not ontological levels.

    If they say, thought, all properties are fundamental, does that mean the property of being a mountain is as ‘fundamental’ as being a fermion, or do mountains not exist in their ontology?

  7. gualtiero piccinini

    In their ontology, mountains are substances and they exist.  However, they are made out of fundamental particles.  So far, this is pretty uncontroversial.  Their view about properties is more controversial.  They deny that any of the ordinary properties (e.g., being snowy, steep, etc.) ascribed to mountains exist.  They think it may be true that mountains are snowy and steep, but it doesn’t follow that there are such properties as snowyness and steepness.  The truthmakers for claims about mountains are to be found among the fundamental properties of the ultimate constituents of mountains.

  8. gualtiero piccinini

    I think it’s fair to say they are reductionist, provided that “reductionism” is defined appropriately.  Ordinary ontological reductionism holds that higher levels reduce to lower levels.  But Heil and Martin don’t think there are any higher levels.  If there are no higher levels, a fortiori there are no reductions of higher levels to lower levels.  Of course, from the point of view of an anti-reductionist, this may sound even more “reductionist” than ordinary reductionism.  But we need to redefine “reductionism” in a new way, e.g., as a view that opposes autonomous higher levels (of being).

  9. Perhaps not the right post to ask, but anyway… Ken, what is your claim exactly? If it is that the existence of a momentary conscious state does not require external scaffolding (except the scaffolding which provides essential support for the brain’s existence) does anyone actually deny that? If it is that a conscious state can’t supervene on external scaffolding, then how does the argument show that?

  10. kenneth aizawa

    “If it is that the existence of a momentary conscious state does not
    require external scaffolding (except the scaffolding which provides
    essential support for the brain’s existence) does anyone actually deny
    that?”

    I think Leibniz denied this in section 7 of the Monadology.  Other than that, I don’t know of anyone who denies this.

    Still, I’m not sure exactly what you mean by “scaffolding.”  Of course, everyone thinks the body provides a heart that pumps nourishing blood to the brain.  Of course, everyone thinks that normal sunlight cast on a tree that your head is oriented toward with eyes open and normal nerves influences what you are conscious of.  The body and environment has a causal role in influencing what one is conscious of.

    “If it is that a conscious state can’t supervene on external
    scaffolding, then how does the argument show that?”

    I’m not claiming that a conscious state can’t supervene on external scaffolding (provided I’ve understood scaffolding the right way).  It is just that as a matter of contingent empirical fact it does not.

    My claim is that there is a defeasible reason to think that conscious states determined by (supervene on, emerge from, are realized by, are determined by, are constituted by) neural states.   The parentheticals are supposed to be variants on the more generic idea of determination.  What is the argument?  Lower-level states determine higher-level states.  Distinct combinations of lower-level properties and relations give rise to distinct higher-level properties, relations, and states.  So, if you have a distinctive higher-level state like consciousness, it should be determined by distinctive lower-level properties and relations?  What meets this last bill?  Neuronal states.

    Here is an intuitive version.  You look at liver cells, brain cells, and muscle cells.  It is perfectly common to suppose that those different types of cells do different things.  So, maybe liver cells metabolize, muscle cells contract, brain cells give rise to consciousness.  This is just ordinary defeasible scientific reasoning as far as I can tell.

  11. kenneth aizawa

    On second thought, let me take back that stuff about Leibniz.  I’m just not that sure of what you are getting at with the quoted passage.

    Here is the specific claim: momentary conscious states are not realized by combinations of brain and bodily states.  Momentary conscious states are realized by brain states.

  12. kenneth aizawa

    Eric,

    The issue of possible top-down effects might be more complicated than
    commentators are allowing. I think Bechtel uses an example of
    explaining the change in position of a molecule by appeal to ethology
    (e.g., the chlorophyl molecule at the tip of the venus fly trap moved
    because a fly landed on the appropriate trigger in the plant). I
    realize this example is uninteresting, but might we be able to pump it
    for ideas?

    Regarding this point, I think a distinction might be in order.  It is one thing for a high-level event, say, the explosion of a bomb to cause a low-level event, such as the depolarization of a nerve cell.  That seems ok.  It is another to have a high-level property, like a diamond’s being hard, to determine a set of low-level properties/relations like the carbon atoms being bonded together in such and such a way.  The latter case seems impossible in the way that, although the length of a pendulum will determine its period, the period of a pendulum cannot determine the length of the pendulum. 

    Bechtel’s kind of case seems to be of the high-level event kind.  That’s, I would have thought, not that interesting.

  13. Flora Carpenter

    While I don’t want to distract the conversation from its more interesting turns, I’d like to clarify as I think I have been sloppy in presenting my view.

    When I said, “parts cannot be considered isolated from the whole,” I meant “knowing the state of parts that comprise a system cannot be considered an exhaustive account of that system as a whole.” And the reason is that I think high-level features have causal properties that effect the system, including arrangement of parts. When I said that properties of high-level states are not determined by their parts, I grossly mispoke (big whoops!). I do believe that the properties of high-level states are determined by the parts and relations among parts. If the properties that emerge only at higher-level states causally impact the direction of change in the system, then it makes sense to include them when describing that system. And this would mean that a description that only considers bottom-up determinism is insufficient. As I see it, that’s one important difference between a brain and a diamond.

    If I am confusing the point or showing symptoms of vitalism, I sincerely apologize.

  14. Ken,

    Yes, I know you believe that conscious states don’t supervene (emrege from…) on external scaffolding. What I wanted to know is how the dream example was supposed to be an argument for that claim, rather than simply a description of one conscious state of which it is true. As for the lower level/higher level claim, that’s neutral between rival accounts of conscious states. What I want is an argument for why *only* neuronal states can play the role of supervenience base. Telling me that *something* has to play the role of supervenience base doesn’t tell me that.

    BTW, brain guys… these threads are hard to read. Responses look like quotes rather than new contributions.

  15. kenneth aizawa

    “Yes, I know you believe that conscious states don’t supervene (emrege
    from…) on external scaffolding. “
        Ok.  So far, so good.  We agree.

    “What I wanted to know is how the
    dream example was supposed to be an argument for that claim, rather
    than simply a description of one conscious state of which it is true.

        I wasn’t intending to advance the dream example as my own.  That was another line that I wished to contrast         with  my own.  But, as you say, the dream example is one conscious state which appears not to depend on the         body.  The argument, thereafter, as I understand it, is a matter of some kind of extrapolation.  Roughly, if the             brain  alone suffices in dreams, why not in general?  So, that’s roughly how the dream argument is supposed to         work.

    “As for the lower level/higher level claim, that’s neutral between rival
    accounts of conscious states. What I want is an argument for why *only*
    neuronal states can play the role of supervenience base. Telling me
    that *something* has to play the role of supervenience base doesn’t
    tell me that.’
        If you want an argument for why *only* neuronal states can play the
    role of supervenience base for conscious         states, I’m not going
    to give one for that.  I don’t believe that.  I argue that only
    neuronal states do play the role     of supervenience base.  The
    reason, I claim, is that neuronal states have a kind of organization
    that is different         from the organization of other bodily
    states.  This difference in organization provides a defeasible reason
    to think        it provides a supervenience base for consciousness
    that, say, muscles do not.  You are exactly right that the idea    
    that there must be a supervenience base for consciousness doesn’t show
    that the brain must be the                         supervenience base. 
    The additional part is that distinctive high-level states must be
    matched by distinctive lwer-level properties/relations.  The brain,
    I’m proposing, provides those distinctive lower-level
    properties/relations.

  16. anna-mari(at)helsinki.fi

    First of all, I`d like to add that there were some non-vitalist British Emergentists. For example C.D Broad offered a non-vitalist notion of emergence (in 1919 and 1925). According to Broad the systemic properties are emergent if they cannot be deduced from the behavior the system`s component.

    As Boogerd et alt emphasize in their article “Emergence and it`s place in Nature” (Synthese, 2000) it is important to distinguish the weak and the strong version of emergentism.

    The weak version is combination of these: The thesis (i) of physical monism, (ii) of organizational properties and (iii) of synchronic determinism.(i) denies the vitalism, since it denies that there would be some supernatural or vitalistic entities whatsoever.

    The (ii) thesis is the source of “weakness”. It specifies the relation between a system`s microstructure and it`s systemic properties. According to it there can be no difference in the systemic properties without there being some differences in the properties of the system`s parts.

    The strong version holds basicly that if there are such properties in some system that cannot be deduced from the components parts, they are called “strongly emergent properties”, since there can be differences in their properties that cannot be deduced from the microlevel.

    However, in Boogerd & co-article is something interesting: B&co distinguish also the vertical and horizontal conditions for emergentism. They are independent of each other.

    The vertical condition is this: A systemic property is emergent, if it is not MECHANISTICALLY explainable. This means that the micro-structural base will not be sufficient to deduce the systemic properties.
    The horizontal condition is this: A systemic property is emergent, if the properties of the parts within the system cannot be deduced from their properties in isolation. This means that for example the properties of part A in the context of system (A,B) are emergent if the were not deducible from properties of A and B in other systems. This means that the behavior of the parts in the system could not be predicted from their behavior in other constellations.

    Now, this is interesting: Boogerd and co argue that for example in cell biology there is emergence even if the emergentist systems or properties can be explained MECHANISTICALLY (in bechtelian-craverian-style). They claim that the source of emergence is the horizontal, not the vertical condition.

    Basicly Boogerd et alt thus argue that even the stronger version of
    emergentism is compatible with the mechanistic model of explanation advocated by Bechtel, Craven and many others, since the source of emergence is not the epistemological part, but the unpredictability.

    This is a nice move, and the paper is quite interesting. However, as Gualtiero would put it, ” I disagree and have objections”:). But since the space-limit, I will leave my objections to rest in peace.
    However, I`d love to hear your opinions.

  17. Ken Aizawa

    Flora,

    My argument began with the premise that higher-level states are determined by lower-level properties and relations. If we agree on that, then that’s one objection to the argument out of the way.

    There are other issues in this area, of course, such as the causal efficacy of higher-level properties, but I won’t go there in this post.

Back to Top
%d bloggers like this: