Lurz, Robert - In Defense of Wordless Thoughts About Thoughts

12/07/2007

Mind & Language Vol 22 No 3 June 2007

This is a paper almost exclusively aimed at refuting Bermudez's theory of nonlinguistic creatures and their capabilities. On Bermudez's account, nonlinguistic creatures can think about the world and have 'protocausal' reasoning, but cannot think thoughts about thoughts, that is, understand that their thoughts stand in relation to themselves-- that is, have 'higher-order propositional attitudes' or higher-order PAs. Author wants to deny this a priori theory.

First, author attacks the a priori aspect of Bermudez's theory by pointing to some empirical work that underlies the position that nonlinguistic animals can entertain higher-order PAs. (pg 272-3) Important to note that Bermudez does not deny that nonlinguistic animals can have thoughts about mental states, just not PAs. (pg 273)

Author quotes Bermudez and describes his theory:
P1) Ascribing PAs involves higher-order thinking (intentional ascent)
P2) Higher-order thinking (intentional ascent) can only be done by using words for the thoughts-- by using public-language sentences. Intentional ascent involves semantic ascent.
Conclusion: PA ascriptions involve public language. (pg 275)

higher-order thinking is considered by Bermudez to be 'consciously considering thoughts and how they relate to each other', which he names 'second-order cognitive dynamics' (pg 276). A point of interest is that this cannot be a sub-personal representation, since then it would conflict with the language of thought hypothesis. Author denies that ascribing PAs need to be done consciously, since there are many studies that seem to show that when children over 4 ascribe PAs to others, they aren't doing it in an acessible manner. (pg 277-9) Author offers a way out for Bermudez by saying that nonlinguistic animals can't explicitly engage in PA ascriptions, but author considers this a much weaker conclusion (pg 280-2).

Author considers an interpretation of Bermudez's theory: call 'first-order cognitive dynamics' the ability to explicitly, reflectively reason about states of affairs (not thoughts). It seems that Bermudez is committed to nonlinguistic animals doing some form of reasoning, but is it 'first-order cognitive dynamics'? If Bermudez says 'yes', then author tries to trap him into admitting that 'first-order' requires language just as much as 'second-order' does. If Bermudez says 'no': 'this happens on the subpersonal level', why wouldn't this happen for 'second-order' as well? (pg 282-4)

Author considers a possible story that suggests that nonlinguistic animals could reason about PAs. (pg 286-7) The story is about a nonlinguistic animal 'tricking' or letting another have a false belief in order to secure a means of amusement for itself. If the story is possible, then it seems that language isn't required to have reasoning about PAs. This goes directly against Bermudez, who thinks that for a PA to be thought about, it must first be represented, and thus requires a 'vehicle' at the 'personal level' (not sub-personal). If they weren't conscious (personal), then we wouldn't be able to 'regulate and police' our thoughts and what we're entitled to believe. The only viable systems available for all of this is analogue representative models (maps, images, models) and language. But Bermudez says that analogue is out (pg 288) so language is the only contender left standing.

Author denies that these two options are the only way to go, and furthermore that Bermudez is confused. He confuses what is a conscious consideration 'the thought' with whatever does the representative work 'the vehicle'. What is at the personal level are thoughts, the vehicles that represent could easily be at the subpersonal. (pg 288-9) Bermudez's apparent reply is that this seems unlikely since whenever you go and check your thoughts, you get words. (pg 289 bottom) Author denies this: there are thoughts that don't come in words, even though you can put them into words (pg 290-1).

Author concludes that it is an empirical issue, not a conceptual one, whether nonlinguistic animals can ascribe PAs.