Short abstracts of some articles that I am currently working on. For full drafts, please send me an email. My email address is my first name followed by my surname followed by @outlook.com. (I am not sure if this (now) common trick will keep a data mining algorithm from figuring out my email address, though.)

[What’s Left of Rigidity?]

This article argues that given some plausible assumptions about natural language and philosophical methodology, the following theses about names cannot be true together: (a) the same (‘proper’) name can be borne by distinct objects/individuals, and (b) names of natural language are rigid designators but definite descriptions are not. The main assumptions that the argument appeals to are: (i) fragments of language (e.g., names) do not themselves refer, but can be used to refer, and (ii) every assumption involved in assessing whether names are (or are not) rigid designators must be preserved in the corresponding assessment for definite descriptions.

[Unspeakable Names, Unwritable Names]

Some names seem to have the same written form but distinct spoken forms e.g., names spelled ‘Jean’ but pronounced differently as /ʤɪ́jn/ and /ʒɑ̃/. It has recently been argued that such cases — call them ‘Gray Cases’ — raise questions concerning name-individuation that are particularly pressing for the predicate view of names (Gray, 2015; Michaelson, 2023; Stojnić, 2023). I argue that the semantic significance of Gray Cases relies on the widely-endorsed and seemingly innocuous — but ultimately false and problematic — metaphysical assumption that the same name can have distinct ‘forms’ e.g., written, spoken, gestural, etc. I argue first, that this assumption does not accord with our usual practice of assigning names (i.e., individuals are assigned sounds, symbols, etc. as names) and the fact that the exact manner of pronouncing/writing a person’s name is important for its identity/individuation, and second, that qua words, names have at most one form.

[Method-Selection Ability and Its Implications for General Intelligence]

In this paper we introduce a formal description of one of a human ability necessary for general intelligence. Then we use this description to argue that it is not possible to create an Artificial General Intelligence, within the current Computer Science paradigm. The ability in question is the ability to identify what method to use for solving some task, and we create a minimally necessary set of variables to represent the ability. Then we show that letting any one of the variables be fully general, fully unrestricted, leads to infinity. This shows that while specialized AI is possible, anything resembling general AI will be either non-autonomous, biased, or non-computable.