Twin Earth, Dry Earth, and Brains in Vats

 

William S. Larkin

Southern Illinois University Edwardsville

 

2003 Meeting of the Central Division of the American Philosophical Association

4/26/03

______________________________________

 

I.                     Introduction

A.      Definitions

1.       BIV Skeptical Argument

Let a BIV be a brain in a vat in an otherwise empty world, hooked up to a super-computer producing for the brain thoughts and experiences that are (narrowly, intrinsically) indistinguishable from my actual thoughts and experiences.

 

SK1: If I know P, then I know that I am not a BIV

SK2:  I do not know that I am not a BIV.

SK3:  So I do not know P

 

2.       Warfield Anti-Skeptical Argument

 

W1:  I am thinking that water is wet.

W2:  If I were a BIV, I could not be thinking that water is wet.

W3:  So I am not a BIV.

 

3.        Privileged Access (PA):  We can know the contents of our own occurrent thoughts without empirical investigation of our behavior or environment.

 

4.        Content Externalism (CE):  Which concepts one has, and thus which thoughts one can think, is determined in part by the physical (and/or social) environment to which one is appropriately (causally) related.

 

5.        Non-Empirical Knowledge:  S can know P non-empirically iff S can know P with no more empirical investigation of S’s environment or behavior than is necessary for S to entertain P.

 

6.        Philosophical Significance of Skepticism

a.        An external world skeptical argument SK is philosophically significant iff it can teach us something about the scope or nature of our knowledge of the external world.

b.       SK is philosophically significant iff it forces us to either (a) acknowledge that a significant range of propositions do not have the positive epistemic status we thought they did, or (b) reject/revise some intuitively plausible claim about what is required for knowledge.

 

B.       Warfield’s Claim:  I can know, on the basis of content externalism and privileged access, that I am not a BIV in a way that does not beg the question against the skeptic.

 

C.       My Thesis:  Warfiled’s claim is not correct—there is no non-question-begging refutation of BIV skepticism on the basis of content externalism and privileged access. 

 

 

 

D.      Significance of Thesis

1.        If Warfield’s claim were correct, then we could resist the conclusion of the BIV skeptical argument without rejecting P1.

2.        If we could do that, then we would not be forced to learn something about either the scope or the nature of our knowledge of the external world in general.

3.        So if Warfield’s claim were correct, then BIV skepticism would not be philosophically significant.

 

II.                   Argument ONE

A.      Concessions

1.        If PA, then I can know W-1 non-empirically.

2.        CE and PA are compatible.

3.        CE and PA are both true.

 

B.       Content Externalism and W-2

1.        Warfield Quote:

“I do claim to know that premise 2 is true and I claim to know this on the basis of my “armchair” a priori reflections on the standard a priori externalist thought experiments…for the purposes of my Antiskeptical Argument I need only for my knowledge of premise 2 to be knowledge that is independent of a perceptual investigation of my environment and this is what I claim it is.  Because the thought experiments do provide overwhelming support for externalism I know that premise 2 is true.  And…my appeal to it does not beg the question against the proponent of the Skpetical Argument.”  [Warfield, “A Priori Knowledge of the World” in Skepticism: A Contemporary Reader (Oxford: Oxford University Press, 1999) p. 80]

 

2.        Twin Earth Argument

 

Twin Earth: World that is superficially indistinguishable from Earth, but where the odorless colorless potable liquid that flows in the rivers and faucets is not H2O but XYZ.  Oscar, a life-long denizen of Earth, and Toscar, a lifelong denizen of Twin Earth, are intrinsically identical and use the term ‘water’ in all the same narrowly described ways. 

 

T1:  The only relevant difference between Oscar and Toscar is that they are (causally) related to distinct physical (and/or social) environments.

T2:  Oscar and Toscar refer in thought and language to different stuff with their uses of the term ‘water’.

            2a:  Oscar refers to H2O but not to XYZ.

            2b:  Toscar refers to XYZ but not to H2O.

T3:  So reference in thought or language is determined in part by one’s (causal) relations to a specific type of (physical/social) environment.

T4:  Concepts and meanings determine reference in thought and language respectively.

T5:  So, concepts and meanings are determined in part by one’s (causal) relations to a specific type of (physical/social) environment. 

 

3.        My knowing W2 requires my knowing that the concept I express with my uses of the term ‘water’ is wide—i.e., is an externalistically individuated concept that one can possess only if one is appropriately related to a specific type of environment.

 

 

 

C.       The Twin Earth argument succeeds in showing that my concept water is wide only if:

1.        My uses of ‘water’ express the same concept as Oscar’s uses of ‘water’.

2.        Water (H2O) and twater (XYZ) are superficially indistinguishable with respect to a significant range R of properties.

3.        Water and twater are substantially distinct—they have distinct underlying (essential) properties [e.g., chemical composition].

 

D.       So the Twin Earth argument succeeds in showing that my concept water is wide only if my term ‘water’ refers to a substantial rather than a superficial kind.

E.       I cannot know non-empirically that my term ‘water’ refers to a substantial and not a superficial kind.

1.        Whether water—i.e., the stuff to which I succeed in referring (if any) with my uses of ‘water’—is a substantial kind or a superficial kind is a property of water.  It is not determined by my intentions regarding the use of ‘water’.

2.        What I do succeed in referring to (if anything) may be determined in part by my intentions, but how that stuff is individuated is not.

3.        It is possible to intend to refer to a substantial kind and yet succeed in referring only to a superficial kind.

4.        Dry Earth:  World where there seems to be an odorless, colorless, potable liquid flowing in the rivers and faucets, but there is not.  It is just a complex and completely convincing hallucination that goes forever undiscovered.

5.        I might intend on Dry Earth to refer to some substantial kind with my uses of ‘water’, and yet succeed in referring only to the superficial kind denoted by something like the complex concept odorless, colorless, potable liquid that flows in rivers and faucets (around here).

6.        Whether the stuff to which I succeed in referring with my uses of ‘water’ is a substantial kind or superficial kind depends on whether I am on Earth or Dry Earth.

7.        I cannot know non-empirically whether I am on Earth or Dry Earth. 

 

F.       Argument Summary

L1:  I can know W2 non-empirically on the basis of the Twin Earth thought experiment only if I can know non-empirically on the basis of the Twin Earth argument that my concept water is wide.

L2:  I can know non-empirically on the basis of the Twin Earth argument that my concept water is wide only if I can know non-empirically that water is a substantial and not a superficial kind.

L3:  I cannot know non-empirically that water is a substantial and not a superficial kind.

L4:  So I cannot know W2 non-empirically on the basis of the Twin Earth thought experiment.

 

III.                 Objection and Reply

A.      Boghossian’s Objection:

1.        Quotation:

“...the Twin Earth externalist is committed...to holding that ‘water’ expresses an atomic concept under conditions where it has a non-empty extension...That is one of the presuppositions of the Twin Earth thought experiment.  But, then, how can the very same word, with the very same functional role, express an atomic concept under one set of external conditions and a compound, deompositional concept under another set of external conditions?  A concept’s compositionality is exclusively a function of its internal ‘syntax’, and can’t be contingent upon external circumstances in the way that the present proposal would require.”  (Boghossian, “What the Externalist Can Know A Priori”, Philosophical Issues 9, 1999, p. 208)

 

2.        The content externalist cannot say that my uses of ‘water’ on Dry Earth, which I intend to express a non-empty atomic concept of a substantial kind, instead express an empty compound concept of a superficial kind.

3.        So (the content externalist must say): if one intends a term to express a concept that refers to a substantial kind, then either it succeeds in doing so or it expresses no concept at all.

4.        Since I can know non-empirically how I intend to use a term and that the term is meaningful (i.e., expresses some concept), I can thereby know non-empirically that my term ‘water’ refers to a substantial kind rather than a superficial kind.

 

B.       Reply

1.        I will admit for the sake of argument that one can know non-empirically that a term is in fact meaningful and that one intends it to express a wide concept.

2.        I will also accept for the sake of argument that a single concept cannot be atomic under certain external circumstances but compound under others.

3.        Fortunately, we are not committed to such a possibility.  The externalist is committed only to saying that a certain term, with a certain functional role, expresses an atomic concept under certain external conditions but a distinct compound concept under other conditions.

4.        The externalist can happily admit and will in fact insist that a single term with a single (narrowly described) functional role can nevertheless express distinct concepts.

 

IV.                Argument TWO

A.      Argument ONE shows only that I cannot know W2 non-empirically on the basis of a Twin Earth-style thought experiment.

B.       The following argument shows that I simply cannot know non-empirically, in any way whatsoever, that any of my concepts are externally individuated.  It follows that there is no way to know W2 without begging the question against the skeptic.

 

L5:  To say that a concept is wide is just to say that one cannot possess that concept without being appropriately related to a specific type of external environment.

L6:  So, to know that my concept C is wide requires knowing that I could not have possessed C had I been related to an environment that is relevantly distinct from my actual environment.  That is, I would have to know that there is some possible world where I do not possess C because that world is distinct from the actual world with respect to features relevant to the possession of C in particular.

L7:  I cannot know whether any specific possible world is distinct from the actual world with respect to certain features unless I know how the actual world is with respect tot those features.

L8:  Since C is in fact a wide concept, some of the features relevant to the possession of C in particular will be specific features of my external environment, which I cannot know about without empirical investigation.

L9:  So, I cannot know non-empirically whether any specific possible world is relevantly distinct from the actual world.

L10:  So, I cannot know non-empirically whether C is wide.

 

V.                  Conclusion

A.      One cannot know non-empirically whether any concept is externally individuated.

B.       So one cannot know non-empirically that a BIV cannot be thinking a certain kind of thought.

C.       Thus there is no viable non-question-begging argument against BIV skepticism on the basis of content externalism and privileged access.

D.      So, the philosophical significance of BIV skepticism remains intact.