Logo: to the web site of Uppsala University

According to some contemporary philosophers, call them easy realists, traditional ontological question, such as those concerning the existence of composite objects, properties, events, numbers, (etc.) can be fully resolved, in favor of realism, by inquiring into the semantics of ordinary language.

Easy realism builds on the assumption that ordinary talk of the aforementioned kinds of entities is guaranteed to be true because it directly reflects in the meaning of the involved expressions. In this paper, I provide an argument against this last claim. By pointing to changes in speech-patterns across counterfactual situations where speakers’ beliefs vary, I argue that that ordinary talk of the relevant entities do not reflect facts about meaning, but rather widespread metaphysical beliefs. As a result, contra easy realism, inquiry into the semantics of ordinary language does not suffice to show that such talk is true. My focus is on Thomasson’s version of the view, but I also briefly discuss how the argument might generalize to the wider easy-realist school.

Some eliminativists about ordinary objects respond to Moorean complaints about contradiction with common sense by arguing that one can combine eliminativism about entities of a kind K with the view that ordinary statements quantifying over/referring to Ks are (often) true. I call such views Quasi-eliminativism. In this paper, I develop a two-tiered critique of such views.

The first tier targets a metasemantic assumption underlying Quasi-eliminativism: that in ordinary language, semantic content is fixed by externalist use-presence correlations, such that talk about e.g., tables is made true by whatever worldly facts happen to systematically correlates with table-talk. While I agree that use-presence correlations generate significant metasemantical pressure, I argue that fit with associated descriptions also play a key role. More specifically, that radical failure to fit associated descriptions precludes content fixation. This opens space for cases where, even if talk about Ks systematically correlates certain worldly facts, those facts may radically fail to match the associated descriptions, meaning that K-talk systematically misfires. Whilst this does not show that Quasi-eliminativism about any particular kind of entities is false, it opens space for a new sort of counter argument based around the presence of radical description errors.

The second tier then develops precisely such an argument. Here I focus on Quasi-eliminativism about ordinary physical objects. I argue that the descriptions we associate with such objects ascribe a kind of “substantiality” to them – a feature which can be analyzed in terms of the properties projected unto the manifest world by innate, core cognitive systems responsible for object cognition. I then argue that, given the sort of sparse, eliminativist physics-metaphysics which Quasi-eliminativists tend to favor, these descriptions are radically incorrect. As a result, and contrary to what Quasi-eliminativists say, ordinary talk about e.g. tables, chairs and mountains systematically misfires given such a metaphysics (or, at best, is rendered indeterminate).

Some have argued that liberal ontologies can avoid paying parsimony-costs for non-fundamental entities so long as they are metaphysically grounded in a sparse fundamental ontology. I argue that this is possible only if the grounders carry full metaphysical responsibility for the grounded, by which I mean that the grounded must be fully constituted, explained or otherwise “traceable back” to its grounders. I also argue that grounding, as ordinarily conceived, fails to guarantee full metaphysical responsibility for the grounded in cases of disjunctive grounds (i.e., case where entities have different grounds at different worlds/times/places) and branching grounds (i.e. cases where one entity grounds two non-identical entities that do not stand in any grounding relation to each other). I explore how the orthodox view on grounding might be revised to overcome this issue, but argue that any revision will end up expanding the fundamental domain with facts about general metaphysical principles/laws. I conclude that non-fundamental entities can be made free of theoretical costs only at the expense of taking on added costs at the fundamental domain.

When an object undergoes sufficiently radical, gradual change, it sometimes appears indeterminate whether the resulting object is numerically identical to the original. Consequently, it seems indeterminate whether the original object still exists. (Think ship of Theseus-style cases.) The “standard” view is to treat such indeterminacy as being de dicto, i.e., as tracing back to the application of referential terms and concepts to post-change objects (which, themselves, determinately exist). In this paper, I develop an alternative account which treats the indeterminacy as being de re, i.e., as tracing back to the existence of the objects themselves.

The key idea of the paper is to analyze statement of the form “it is indeterminate whether A exists” as claims about a past object, and asserting that it is indeterminate whether there is any object now (or in the future) that is identical to it. I demonstrate that this analysis, if construed correctly, dodges traditional criticisms of de re ontological indeterminacy, such as that it is incoherent, leads to issues with truth and reference, and violates the binary nature of existence. In so doing, the paper effectively demystifies the notion or de re ontological indeterminacy, putting it back on the philosophical table. In addition, I also argue that this de re analysis enjoys certain theoretical advantages over traditional de dicto views, including compatibility with leaner ontologies, and more intuitive views about the referential properties of proper names and pronouns, as well as mapping better to the contents expressed in ordinary speech and thought about cases of gradual change.

In this paper, I present a response to Gareth Evan’s (in)famous argument against the possibility of indeterminate identity. Evans argues that statements of the form “it is indeterminate whether a = b” lead to contradiction. I argue that, once we’ve taken into account (i) how properties must distribute across indeterminately identical objects, and (ii) how properties that include indeterminately identical objects in their definition behave logically, we see that a crucial step in Evans’ argument is invalid. More specifically, I argue that, once we take these facts into account, we can see that, if it is indeterminate whether a = b, then it is not the case, as Evans’ claims, that b has the property of being indeterminately identical to a. Instead, b only indeterminately has this property, which blocks the derivation of Evans’ contradiction.