Abstract
The connectives of classical propositional logic are given an analysis in terms of necessary and sufficient conditions of acceptance and rejection, i.e. the connectives are analyzed within an expressivist bilateral meaning-is-use framework. It is explained how such a framework differs from standard (bilateral) inferentialist frameworks and it is argued that it is better suited to address the particular issues raised by the expressivist thesis that the meaning of a sentence is determined by the mental state that it is conventionally used to express. Furthermore, it is shown that the classical requirements governing the connectives completely characterize classical logic, are conservative (indeed make the connectives redundant) and separable, are in bilateral harmony, are structurally preservative with respect to the classical coordination requirements and resolve the categoricity problem. These results are taken to show that one can give an expressivist bilateral meaning-is-use analysis of the connectives that confer on them a determinate coherent classical interpretation.
Similar content being viewed by others
Notes
I am ignoring throughout all the complications that come with sentences containing indexical or deictic elements.
To see the problem, note that on the classical interpretation of \(\exists \), rejecting every instance of \(A(a)\) (where \(a\) is an individual constant) should be compatible both with rejecting \(\exists x A(x)\) and with accepting \(\exists x A(x)\). One’s attitudes of acceptance and rejection towards non-quantified sentences are just not in general sufficient to determine one’s attitudes of acceptance and rejection towards quantified sentences.
One can resist this conclusion. One can take the expressive requirements governing the connectives to express the meaning, or cognitive content, of the connectives: to interpret a sentence is to break it down step by step into its basic components entering (perhaps provisionally) into the mental states specified by the expressive rules; this procedure gives the sentence its cognitive meaning (this is reminiscent of Kracht 2011). The expressive requirements and not the structural requirements would then be constitutive of the meaning of the connectives and the contribution of the structural requirements to logic would be ‘extra-semantic’ or epistemic. The result would be a conception of logic in which some logical truths and entailments (such as the law of excluded middle and modus ponens) would be a priori yet would not be analytic. This is closely related to Tennant’s (e.g. 2014) idea of a ‘core logic’. However, Tennant’s core logic is developed along the lines of traditional inferentialism rather than expressivism and this results in a stronger logic. The core logic that one would get if one discounted all the structural requirements and maintained only the expressive requirements would be very weak, comprising, in the present language, only the de Morgan equivalences (e.g. \(A\) would be equivalent to \(\lnot \lnot A\), but from \(A\) and \(A\supset B\) it would not follow that \(B\)).
References
Belnap, N. D. (1962). Tonk, Plonk and Plink. Analysis, 22(6), 130–134.
Bendall, K. (1979). Negation as a sign of negative judgement. Notre Dame Journal of Formal Logic, 20, 68–76.
Blackburn, S. (1984). Spreading the word. Oxford: Oxford University Press.
Blackburn, S. (1988). Attitudes and contents. Ethics, 98(3), 501–517.
Brandom, R. (1994). Making it explicit. Cambridge: Harvard University Press.
Brandom, R. (2008). Between saying and doing. Oxford: Oxford University Press.
Cantwell, J. (2013). Expressivist first order logic. Erkenntnis, 78(6), 1381–1403. doi:10.1007/s10670-012-9421-4.
Cantwell, J. (2014). Unity and autonomy in expressivist logic. Dialectica, 68, 443–457.
Carnap, R. (1937). Philosophy and logical syntax. London: Kegan Paul, Trench, Trubner.
Carnap, R. (1943). Formalisation of logic. Cambridge, MA: Harvard University Press.
Chrisman, M. (2008). Expressivism, inferentialism, and saving the debate. Philosophy and Phenomenological Research, 77(2), 334–358.
Dreier, J. (2006). Negation for expressivists: A collection of problems with a suggestion for their solution. In R. Shafer-Landau (Ed.), Oxford studies in metaethics (Vol. 1, pp. 217–234). Oxford: Oxford University Press.
Dummett, M. (1991). The logical basis of metaphysics. London: Duckworth.
Ferreira, F. (2008). The co-ordination principles: A problem for bilateralism. Mind, 117, 1051–1057.
Gentzen, G. (1969). Investigations into logical deduction. In M. Szabo (Ed.), The collected works of Gerhard Gentzen. Amsterdam: North Holland.
Goldman, A. I. (2006). Simulating minds: The philosophy, psychology, and neuroscience of mindreading. Oxford: Oxford University Press.
Hacking, I. (1979). What is logic? The Journal of Philosophy, 76, 285–319.
Hale, B. (1993). Can there be a logic of attitudes? In J. Haldance & C. Wright (Eds.), Reality, representation, and projection. New York: Oxford University Press.
Kracht, M. (2011). Gnosis. Journal of Philosophical Logic, 40, 397–420.
Levi, I. (1996). For the sake of the argument. Cambridge: Cambridge University Press.
Prawitz, D. (1965). Natural deduction: A proof-theoretical study. Stockholm: Almqvist & Wiksell.
Price, H. (1983). Sense, assertion, dummett and denial. Mind, 92, 161–173.
Price, H. (1990). Why ‘not’? Mind, 99(394), 221–238.
Prior, A. (1960). The runabout inference ticket. Analysis, 21(2), 38–39.
Rumfitt, I. (2000). Yes and no. Mind, 109, 781–823.
Rumfitt, I. (2008). Co-ordination principles: A reply. Mind, 117, 1059–1063.
Sandqvist, T. (2009). Classical logic without bivalence. Analysis, 69(2), 211–218.
Schroeder, M. (2008a). Being for: Evaluating the semantic program of expressivism. Oxford: Oxford University Press.
Schroeder, M. (2008b). Expression for expressivists. Philosophy and Phenomenological Research, 76(1), 86–116.
Shoesmith, D. J., & Smiley, T. J. (1978). Multiple conclusion logic. Cambridge: Cambridge University Press.
Smiley, T. (1996). Rejection. Analysis, 56, 1–9.
Steinberger, F. (2011). What harmony could and could not be. Australasian Journal of Philosophy, 89(4), 617–639.
Tennant, N. (1999). Negation, absurdity and contrariety. In D. Gabbay & H. Wansing (Eds.), What is negation? (pp. 199–222). Dordrecht: Kluwer.
Tennant, N. (2014). Logic, mathematics and the a priori, part ii: Core logic as analytic, and as the basis for natural logicism. Philosophia Mathematica. doi:10.1093/philmat/nku009.
Author information
Authors and Affiliations
Corresponding author
Technical Appendix I: Nelson Models
Technical Appendix I: Nelson Models
Most of the central results of the paper (proved in the next section) utilize the close relation between accept–reject states and Nelson models defined in this section. Nelson models are used for their technical power, not because they are taken to provide an intuitively satisfactory rendition of the basic notions of acceptance and rejection.
A Nelson model is a structure \(N = (I, \sqsubseteq , V^{+}, V^{-})\) such that (i) \(I\) is a set, (ii) \(\sqsubseteq \) partially orders \(I\), (iii) for each \(p\) in the base language \(V^{+}(p), V^{-}(p) \subseteq I\), and (iv) satisfies persistence (\(\circ \in \{+,-\}\)):
On the basis of such a model one can define a positive and negative forcing relation \(\models \), as follows:
The accept–reject state \(\Vdash \) induced by a Nelson model \(N\) (with corresponding forcing relation \(\models \)) is defined (for any finite sets \(\Gamma \) and \(\Delta \) of sentences from the extended language and any \(A\) from the extended language):
Notation: \(+\Gamma = \{+A\,|\, A\in \Gamma \}\) and \(u\models ^{+}\Gamma \) is a short-hand for \(\forall A\in \Gamma : u\models ^{+}A\).
I state here, without proofs (they employ standard techniques), some properties that will be employed later:
Lemma 1
-
1.
The accept–reject state \(\Vdash \) induced by \(N\) satisfies, Reflexivity, Monotonicity and Cut, and the classical expressivist requirements for \(\lnot \) and \(\supset \).
-
2.
Let \(\Vdash \) be an accept–reject state induced by a Nelson model \(N\). If there is no \(u\) and \(p\) such that \(u\in V^{+}(p)\) and \(u\in V^{-}(p)\), then \(\Vdash \) satisfies Explosion.
-
3.
Let \(\Vdash \) be an accept–reject state induced by a Nelson model \(N\). \(\Vdash \) will satisfy Classical Reductio if \(\models \) satisfies (for all \(u, v\in I\) and atomic letters \(p\)):
-
(a)
If there is no \(v\) such that \(u\sqsubseteq v\) and \(v\in V^{-}(p)\), then \(u\in V^{+}(p)\).
-
(b)
If there is no \(v\) such that \(u\sqsubseteq v\) and \(v\in V^{+}(p)\), then \(u\in V^{-}(p)\).
-
(a)
The canonical model for \(\Vdash \) is the Nelson model \(N = (I, V^{+}, V^{-}, \sqsubseteq )\). Here \(I\) is the set of suppositional contexts (so when \(u\in I\), \(u\) will be some set \(\{+A_1, \ldots , +A_n, -B_1, \ldots , -B_m\}\)), with the restriction that for some sentence \(A\), either \(u\not \Vdash +A\) or \(u\not \Vdash -A\). Furthermore, \(V^{+}\) and \(V^{-}\) are defined:
Finally, \(\sqsubseteq \) is defined:
One can show using standard techniques that the canonical model for an accept–reject state induces that accept–reject state (so every accept–reject state is induced by some Nelson model).
1.1 Proof of Main Theorems
The proofs of Theorems 1 and 2 are straightforward.
Proof of Theorem 3
The proof idea is to use \(\Vdash _b\) to construct a Nelson model and use such a model to construct an accept–reject state \(\Vdash _e\) and then show that \(\Vdash _e\) is a conservative extension of \(\Vdash _b\) with the relevant properties.
Let \(I\) be the set of suppositional contexts \(u\) (restricted to the base language) such that for some \(p\), either \(u\not \Vdash _b +p\) or \(u\not \Vdash _b - p\).
Define:
Define:
\(\sqsubseteq \) is a partial order. Through Reflexivity, \(u\sqsubseteq u\) for all \(u\). Assume that \(u\sqsubseteq v\) and \(v\sqsubseteq w\). So \(\forall p\in u^{+}\): \(v\Vdash _b + p\), and (ii) \(\forall p\in u^{-}\): \(v\Vdash _b - p\). Furthermore, \(\forall p\in v^{+}\): \(w\Vdash _b + p\), and (ii) \(\forall p\in v^{-}\): \(w\Vdash _b - p\). Assume that \(p\in u^{+}\). Then \(v\Vdash _b + p\). By Monotonicity, \(v\cup w \Vdash _b + p\). By repeated applications of Cut, \( w \Vdash _b + p\). Similarly, if \(p\in u^{-}\), then \( w \Vdash _b - p\). So \(u \sqsubseteq w\).
\(\sqsubseteq \) is persistent. Assume that \(p\in u^{+}\) and \(u\sqsubseteq v\). Then \(v\Vdash _b + p\), so then \(v\in V^{+}(p)\). Assume that \(p\in u^{-}\) and \(u\sqsubseteq v\). Then \(v\Vdash _b - p\), so then \(v\in V^{-}(p)\).
So we have a Nelson model.
Let \(\Vdash _e\) be the accept–reject state induced by the forcing relation \(\models \) in the above Nelson model.
One needs to show that \(\Vdash _e\) is identical to \(\Vdash _b\) when restricted to the base language. So let \(\Gamma = \{p_1, \ldots , p_m\}\) and \(\Delta = \{q_1, \ldots , q_m\}\) be sets of atomic sentences. \(+p_1, \ldots , +p_n, -q_1, \ldots , -q_m \Vdash _e + r\) if and only if \(\forall v:\) if \(v \models ^{+}p_i\) (for each \(i\)) and \(v\models ^{-}q_j\) (for each \(j\)), then \(v \models ^{+}r\) if and only if \(\forall v:\) if \(v \in V^{+}(p_i)\) (for each \(i\)) and \(v\in V^{+}(q_j)\) (for each \(j\)), then \(v \in V^{+}(r)\) if and only if \(\forall v:\) if \(v\Vdash _b + p_i\) (for each \(i\)) and \(v\Vdash _b - q_j\) (for each \(j\)), then \(v \Vdash _b + r\) if and only if (given Monotonicity and Cut) \(+p_1, \ldots , +p_n, -q_1, \ldots , -q_m \Vdash _b + r\).
So \(\Vdash _e\) is a conservative extension of \(\Vdash _b\) that (Lemma 1) satisfies the minimal structural requirements and the classical expressivist requirements for the connectives.
Finally, we need to show that if \(\Vdash _b\) satisfies Explosion, then so does \(\Vdash _e\) and if \(\Vdash _b\) satisfies Classical Reductio, then so does \(\Vdash _e\).
Explosion If \(\Vdash _b\) satisfies Explosion, then there is no \(u\) and no \(p\) such that \(u\models ^{+}p\) and \(u\models ^{-}p\). By Lemma 1, \(\Vdash _e\) satisfies Explosion.
Classical Reductio Assume that \(\Vdash _b\) satisfies Classical Reductio. First establish the following two claims:
-
1.
If there is no \(v\) such that \(u\sqsubseteq v\) and \(v\in V^{-}(p)\), then \(u\in V^{+}(p)\).
-
2.
If there is no \(v\) such that \(u\sqsubseteq v\) and \(v\in V^{+}(p)\), then \(u\in V^{-}(p)\).
Assume that \(u\not \models ^{+}p\). Then \(u\not \in V^{+}(p)\) and so \(u\not \Vdash +p\). By Classical Reductio, \(u \cup \{ - p\}\not \Vdash +p\). But then \(u \cup \{ - p\}\in I\) and as \(u \cup \{ - p\}\in V^{-}(p)\), \(u \cup \{ - p\} \models ^{-}p\). Furthermore, \(u \sqsubseteq u \cup \{ - p\}\). So there is a \(v\) such that \(u\sqsubseteq v\) and \(v\models ^{-}p\). The second case is analogous.
This establishes (1) and (2) above. From Lemma 1 it follows that \(\Vdash _e\) satisfies Classical Reductio.
Proof of Theorem 4
Notation: where \(\mathbf{S}= \{+A_1, \ldots , + A_n, -B_1, \ldots , -B_m\}\), let \(S_\mathbf{S}= A_1 \wedge \cdots \wedge A_n \wedge B_1 \wedge \cdots \wedge B_m\).
Note that any sentence \(A\) is classically equivalent to a (finite) set \(\Gamma \) of sentences of either the form \(p_1 \wedge \cdots \wedge p_n \wedge \lnot q_1\wedge \cdots \wedge \lnot q_m \supset r\) or the form \(p_1 \wedge \cdots \wedge p_n \wedge \lnot q_1\wedge \cdots \wedge \lnot q_m \supset \lnot r\) (that is \(A\) is a classical consequence of such a set \(\Gamma \), and every element of \(\Gamma \) is a consequence of \(A\)). Thus (from Theorem 2) where \(\Gamma \) is such a set, a classical accept–reject state will satisfy:
A classical accept–reject state \(\Vdash _e\) defined on the extended language is uniquely determined by the set of categorically accepted sentences (Theorem 2). So as acceptance of a categorical sentence in turn is uniquely determined by the restriction of \(\Vdash \) to the base language, it follows that as (Theorem 3) every classical accept–reject state \(\Vdash _b\) defined on the base language has a classical conservative extension defined on the extended language, it has precisely one such extension.
Proof of Theorem 5
(1) Assume that \(\Vdash \) satisfies the classical requirements for accepting \(\lnot \).
Left-to-right, i.e. assume that \(\Vdash \) satisfies the classical requirements for accepting \(\lnot \).
Assume that \(\not \Vdash -\lnot A\). Assume, for reductio, that \(\Vdash + A\). By Reflexivity \(+\lnot A\Vdash + \lnot A\), i.e. \(+\lnot A\Vdash - A\). By Monotonicity \(+\lnot A\Vdash + A\). So, by Explosion, \(+\lnot A\Vdash - \lnot A\). By Classical Reductio, \(\Vdash - \lnot A\). But then we have a contradiction, so \(\not \Vdash +A\).
For the converse direction, Assume that \(\not \Vdash +A\). Assume for reductio that \(\Vdash - \lnot A\). By Reflexivity \(- A\Vdash - A\), i.e. \(-A \Vdash + \lnot A\). By Monotonicity \(- A\Vdash - \lnot A\). By Explosion \(-A\Vdash + A\), so by Classical Reductio \(\Vdash + A\) and we have a contradiction. So \(\not \Vdash -\lnot A\).
Right-to-left, i.e. assume that \(\Vdash \) satisfies the classical requirements for rejecting \(\lnot \). Assume that \(\not \Vdash +\lnot A\). Assume, for reductio, that \(\Vdash - A\). By Reflexivity \(-\lnot A\Vdash - \lnot A\), i.e. \(-\lnot A\Vdash + A\). By Monotonicity \(-\lnot A\Vdash - A\). So, by Explosion, \(-\lnot A\Vdash + \lnot A\). By Classical Reductio, \(\Vdash + \lnot A\). But then we have a contradiction, so \(\not \Vdash -A\).
For the converse direction, Assume that \(\not \Vdash -A\). Assume for reductio that \(\Vdash + \lnot A\). By Reflexivity \(+ A\Vdash + A\), i.e. \(+A \Vdash - \lnot A\). By Monotonicity \(+ A\Vdash + \lnot A\). By Explosion \(+A\Vdash - A\). But then by Classical Reductio \(\Vdash - A\) and we have a contradiction. So \(\not \Vdash +\lnot A\).
(2) Left-to-right, i.e. assume that \(\Vdash \) satisfies the classical requirements for accepting \(\supset \).
Assume that \(\Vdash - A\supset B\). By Reflexivity and Explosion \(- A, +A \Vdash + B\) and so \(-A\Vdash + A\supset B\). By Monotonicity \(-A\Vdash - A\supset B\) and so by Explosion \(-A\Vdash + A\). By Classical Reductio, \(\Vdash + A\). By Reflexivity and Explosion \(+ B, +A \Vdash + B\) and so \(+B\Vdash + A\supset B\). By Monotonicity \(+B\Vdash - A\supset B\) and so by Explosion \(+B\Vdash - B\). By Classical Reductio, \(\Vdash - B\).
Assume that \(\Vdash + A\) and \(\Vdash - B\). By Reflexivity \(+A\supset B\Vdash + A\supset B\), i.e. \(+A\supset B, +A\Vdash + B\). By Monotonicity \(+A\supset B, +A\Vdash - B\) and so by Explosion \(+A\supset B, +A\Vdash - A\). By Reflexivity \(+A\supset B, +A\Vdash + A\) so by Explosion \(+A\supset B, +A \Vdash - A\supset B\). By Monotonicity \(+A\supset B\Vdash + A\) and so by Cut \(+A\supset B\Vdash - A\supset B\). By Classical Reductio, \(\Vdash - A\supset B\).
Right-to-left, i.e. assume that \(\Vdash \) satisfies the classical requirements for rejecting \(\supset \).
Assume that \(\Vdash + A\supset B\). By Reflexivity \(+A, -B\Vdash + A\) and \(+A, -B\Vdash - B\). So \(+A, -B\Vdash - A\supset B\). By Monotonicity \(+A, -B\Vdash + A\supset B\). By Explosion \(+A, -B\Vdash + B\). By Classical Reductio \(+A\Vdash + B\).
Assume that \(+A\Vdash + B\). By Monotonicity \(+A, -A\supset B\Vdash + B\). By Reflexivity \(+A, -A\supset B\Vdash - A\supset B\) and so \(+A,-A\supset B\Vdash - B\). By Explosion \(+A, -A\supset B \Vdash + A\supset B\). By Classical Reductio, \(+A\Vdash + A\supset B\). By Reflexivity, \(-A\supset B \Vdash - A\supset B\). Thus \(-A\supset B\Vdash + A\). By Monotonicity, \(-A\supset B, +A\Vdash + A\supset B\). So, by Cut, \(-A\supset B\Vdash + A\supset B\). By Classical Reductio, \(\Vdash + A\supset B\).
Proof of Theorem 6
Let \(\Vdash \) be an accept–reject state satisfying the minimal structural requirements and the expressive classical requirements on the connectives as well as the restricted requirements Explosion(R) and Classical Reductio(R). Let \(N\) be the canonical model for \(\Vdash \).
For Explosion assume that \(u\in I\) and, for reductio, that \(u\models ^{+}p\) and \(u\models ^{-}p\). It follows that \(u\Vdash + p\) and \(u\Vdash - p\). By Explosion(R) \(u\Vdash + q\) and \(u\Vdash - q\) for all \(q\). But then, this is easy to show, \(u\Vdash + A\) and \(u\Vdash - A\) for all \(A\). But this contradicts the assumption that \(u\in I\). So there is no \(u\in I\) and no \(p\) such that \(u\models ^{+}p\) and \(u\models ^{-}p\). So by Lemma 1, \(\Vdash \) satisfies Explosion.
For Classical Reductio, establish the following two claims:
-
1.
If there is no \(v\) such that \(u\sqsubseteq v\) and \(v\in V^{-}(p)\), then \(u\in V^{+}(p)\).
-
2.
If there is no \(v\) such that \(u\sqsubseteq v\) and \(v\in V^{+}(p)\), then \(u\in V^{-}(p)\).
Assume that \(u\not \in V^{+}(p)\). It follows that \(u\not \Vdash +p\). By Classical Reductio(R) it follows that \(u, -p\not \Vdash +p\). So \(u\cup \{-p\}\in I\). Note that \(u\sqsubseteq u\cup \{-p\}\) and \(u\cup \{-p\}\in V^{-}(p)\). This establishes (1), the proof of (2) is analogous. From Lemma 1 it follows that \(\Vdash \) satisfies Classical Reductio.
Rights and permissions
About this article
Cite this article
Cantwell, J. An Expressivist Bilateral Meaning-is-Use Analysis of Classical Propositional Logic. J of Log Lang and Inf 24, 27–51 (2015). https://doi.org/10.1007/s10849-014-9209-9
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10849-014-9209-9