Mind and paradox: paradoxes depend on minds

Abstract

Paradoxes are mind-dependent in a number of ways. First, by definition, paradoxes offer surprises or apparent contradictions. Since surprise and appearance rely on subjective psychological reactions, paradoxes rely on psychological events. Second, propositional versions of the liar paradox must eventually appeal to sentences if they are to achieve traction, yet sentential versions of the liar paradox rely on language and hence on mentality. Third, belief paradoxes such as B, "No one believes B", transparently hinge on the existence of mental states. (Belief paradoxes extending work by Burge, Buridan, Grim, Sainsbury, and Sorensen demonstrate a fundamental limitation on all cognitive systems: that none can hold all and only truths.) Finally, it is argued that belief paradoxes, like liar paradoxes, arise from the nature of semantic representation.

Keywords:

• incompleteness
• blindspot
• belief paradox
• epistemic paradox
• semantic paradox
• anti-realism

Acknowledgements

My work has benefited by kind assistance from Jesse Prinz, R.A. Lafferty and an anonymous reviewer.

Notes

1. For acute criticisms of possible-world semantics, however, see Forster (2005).

2. Despite initial appearances, the three dimensions identified in the text are independent of each other. (i) It might be thought that if a sentence is an utterance then its structure must be flat. However, anyone who wishes to identify sentences with strings of syllables is likely to want to put a rider on such analysis: syllables produced by random wind, even if they seem to form meaningful patterns, are not quite sentences; it is only syllable sequences accompanied by suitable intentions that count as genuine sentences. By linking syllables to intentions, we can thereby assign the syllables hierarchical structure, if we so wish. (ii) Conversely, it might be thought that if a sentence is a mental representation then its structure must be hierarchical. While it is true that some mental representations will be bracketed, others will not: in the initial stage of audition, and the final stage of motor production, our mental representations of speech will be flat. (iii) It might be thought that if a sentence is an abstract object then it must be a type rather than a token. However, this is not the case. True, the main rationale for positing abstract objects has been to explicate or to ground the type–token relation; and true, Platonism is naturally regarded as positing abstract objects as types that subsume physical objects as tokens. Yet, there is no contradiction in saying that Plato's heaven includes two numerically distinct Forms belonging to a single type of a given expression (perhaps a Platonic form at a higher realm of abstraction, perhaps a type understood in nominalist terms).

3. Why would we pretend that mindless marks possess significance? Once I read a book of computer-generated poetry just to see how much sensible-seeming content could be produced by stochastic means. Others may have read the book in order to be freed from the shackles of authorial intent, so that they could construct their own meanings without doing the work of arranging their own wordings. Others still may seek mystic significance behind the random words of (say) an I Ching toss – but this latter is not really a case of mindless meaning, for superstitious interpreters believe that they are decoding the messages of invisible forces that have intelligence. Most important of all, we have a natural instinct to treat homophonous elements as words in our own language; we interpret as a matter of reflex, even when we know we should not.

4. For expository convenience, I treat sentences as propositions (i.e. as truth-bearing objects of belief). All of my arguments can be converted into wordier versions premised on alternative views of what propositions are. For details, and significant extensions of Section 4 more generally, see Saka (2012).