Fulltext

Computable Diagonalizations and Turing's Cardinality Paradox

Jacquette, Dale

In: Journal for General Philosophy of Science, 2014, vol. 45, no. 2, p. 239-262

Add to personal list
    Title
    • Computable Diagonalizations and Turing's Cardinality Paradox
    Author
    • Jacquette, Dale. University of Bern, Bern, Switzerland
    Document Type
    • Postprint
    Language
    • English
    Published in
    • Journal for General Philosophy of Science, 2014, vol. 45, no. 2, p. 239-262. Springer Netherlands
    Other Version
    OAI-PMH Identifier
    • oai:doc.rero.ch:326054
    Summary
    A. N. Turing's 1936 concept of computability, computing machines, and computable binary digital sequences, is subject to Turing's Cardinality Paradox. The paradox conjoins two opposed but comparably powerful lines of argument, supporting the propositions that the cardinality of dedicated Turing machines outputting all and only the computable binary digital sequences can only be denumerable, and yet must also be nondenumerable. Turing's objections to a similar kind of diagonalization are answered, and the implications of the paradox for the concept of a Turing machine, computability, computable sequences, and Turing's effort to prove the unsolvability of the Entscheidungsproblem, are explained in light of the paradox. A solution to Turing's Cardinality Paradox is proposed, positing a higher geometrical dimensionality of machine symbol-editing information processing and storage media than is available to canonical Turing machine tapes. The suggestion is to add volume to Turing's discrete two-dimensional machine tape squares, considering them instead as similarly ideally connected massive three-dimensional machine information cells. Three-dimensional computing machine symbol-editing information processing cells, as opposed to Turing's two-dimensional machine tape squares, can take advantage of a denumerably infinite potential for parallel digital sequence computing, by which to accommodate denumerably infinitely many computable diagonalizations. A three-dimensional model of machine information storage and processing cells is recommended on independent grounds as better representing the biological realities of digital information processing isomorphisms in the three-dimensional neural networks of living computers.