You are on page 1of 8

On an inferential semantics for

classical logic
DAVID MAKINSON∗ , Department of Philosophy, Logic and Scientific Method,
London School of Economics, Houghton St, London WC2A 2AE, UK

Abstract

Downloaded from http://jigpal.oxfordjournals.org/ at University of Michigan on June 15, 2015


We seek a better understanding of why an inferential semantics devised by Tor Sandqvist yields full classical logic, by
providing and analysing a direct proof via a suitable maximality construction.

Keywords: Inferentialism, proof-theoretic semantics, production rules, classical logic, maxiconsistency.

1 Introduction
Sandqvist [3, 4] devised a semantics for first-order logic, in which connectives are evaluated on sets
of ‘basic rules’ rather than on assignments of truth-values, and showed that despite its inferential
character it yields full classical logic. It thus provides the first known inferential semantics for
classical logic, upsetting the widely held presumption that such a style of semantics must lead to a
weaker, typically intuitionistic, logic.
As recognized by Sandqvist and emphasized in a proof-theoretic discussion by de Campos Sanz,
Piecha and Schroeder-Heister [1], the system is not very robust in its choice of primitive connectives.
If, like Sandqvist, we take → and ⊥ as primitive for the propositional part, with ¬, ∨, ∧, treated as
abbreviations in a familiar manner, then the semantics yields all of classical logic; but if we treat ∨
as primitive with a classical or Kripke-style evaluation rule, then it becomes sub-classical.
Despite this sensitivity, the construction remains interesting and, moreover, rather puzzling. Why
does it yield classical logic despite its inferentialist character and its similarity to Kripke seman-
tics, which yields only intuitionistic logic? In this note we seek a better understanding of the
situation.
Sandqvist’s original proof is not very helpful in this regard. It proceeds by checking the validity
of the axioms of minimal logic under his semantics, plus the axiom of double negation elimination,
drawing on proof-theoretic results from Prawitz [2] to verify the latter. The desired result then
follows, since those axioms are known to be classically complete.
We give a direct proof with a ‘maxiconsistent basis’ argument that is analogous to, but not
quite the same as, the familiar Lindenbaum/Henkin maxiconsistent set-of-formulas construction.
This gives a clearer picture of the inner mechanisms, and allows easy identification of the places
where proof would snag under various small changes in the evaluation procedure. It also per-
mits one to formulate some variants and extensions of the Sandqvist semantics without loss of
classicality.
The proof is given initially in the propositional context as that allows the clearest view of what is
going on, without distractions. It is then extended routinely to the first-order context.


E-mail: david.makinson@gmail.com

Vol. 22 No. 1, © The Author 2013. Published by Oxford University Press. All rights reserved.
For Permissions, please email: journals.permissions@oup.com
doi:10.1093/jigpal/jzt038 Advance Access published 7 November 2013

[17:18 21/12/2013 jzt038.tex] Paper Size: a4 paper Job: JIGPAL Page: 147 147–154
148 Inferential semantics for classical logic

2 The inferential semantics


By and large we follow Sandqvist’s terminology; exceptions will be signalled. The propositional
language is assumed to have countably many sentence letters, and at most the connectives →, ⊥
as primitive. A basic rule is any pair (Lf ,p) where Lf is a finite (possibly empty) set of sentence
letters and p is a sentence letter; it is conventionally written as p1 ,...,pn ⇒ p where {p1 ,...,pn } = Lf .
In computer science literature, basic rules are also sometimes called production rules.
A basis is any collection (finite or countable) of basic rules. When B is a basis and K is a set of
sentence letters, we write B(K) (with B underlined) for the closure of K under all the rules in B in
the usual set-theoretic sense. That is, B(K) is the intersection of all sets L of sentence letters such
that K ⊆ L and whenever (Lf ,p) ∈ B with Lf ⊆ L then p ∈ L.

Downloaded from http://jigpal.oxfordjournals.org/ at University of Michigan on June 15, 2015


Remarks on closure under bases

(a) Thus for any basis B, the function B(·) is a closure operation (i.e. satisfying inclusion, monotony,
idempotence) taking each set K of sentence letters to a set B(K) of sentence letters.
(b) Clearly, for every basis B the operation B(·) is compact in two distinct senses. It is sentence-
compact in the sense that whenever p ∈ B(K) then p ∈ B(Kf ) for some finite set Kf ⊆ K of
sentence letters. It is also rule-compact, in the sense that whenever p ∈ B(K) then p ∈ Bf (K) for
some finite basis Bf ⊆ B.
(c) Sentence-compactness will not be important for us, for we will never have to consider values of
B(K) for K = ∅. But rule-compactness will be essential for our proof of completeness, which
would snag (at Lemma 3.3) if one allowed it to fail (e.g. by allowing rules with countably many
premises).

Sandqvist’s semantics gives a recursive definition of a satisfaction relation  between bases B on


the left and formulae ϕ on the right. It may be formulated (in the propositional context) as follows.

For sentence letters p: B  p iff p ∈ B(∅).


For ⊥: B  ⊥ iff p ∈ B(∅) for all sentence letters p.
For →: B  ϕ → ψ iff for all C ⊇ B, if C  ϕ then C  ψ.

Remarks on the evaluation rules

(a) Some readers may not be happy with the term ‘semantics’ for this construction, feeling that
it—and perhaps any inferential semantics—is too syntactic in nature to deserve that title. They
may prefer to use the neutral term ‘evaluation system’; none of our formal results depend on
the choice of terminology.
(b) This way of evaluating the falsum has a long history in inferentialist circles, but is quite different
from the one customary in classical logic (or even in Kripke models for intuitionistic logic)
where ⊥ is counted false under every classical valuation (resp. is satisfied in no state in any
Kripke model). In the present context, the analogue of that would be: B  ⊥ for all bases B,
and for future reference we will call that the classical-style evaluation of ⊥. We will see that it
makes a difference which evaluation rule is used.
(c) If one wishes to evaluate sequents as well, one may follow Sandqvist in putting B  (A ⇒ ψ),
where A is a set of formulae and ψ is a formula, iff for all C ⊇ B, if C  ϕ for all ϕ ∈ A then
C  ψ. However, we will not need to consider sequents for our results.

A formula ϕ is called valid in the Sandqvist semantics iff B  ϕ for every basis B. Equivalently (once
one proves Lemma 3.1a below), iff ∅  ϕ.

[17:18 21/12/2013 jzt038.tex] Paper Size: a4 paper Job: JIGPAL Page: 148 147–154
Inferential semantics for classical logic 149

Remarks on validity and the supply of sentence letters

(a) Given Sandqvist’s way of evaluating the falsum, it is important that the language has infinitely
many sentence letters. If there were only n letters p1 ,...,pn then the semantics would val-
idate the non-tautology p1 → (... → (pn → ⊥)...), thus destroying its soundness wrt classical
logic.
(b) At first glance, this feature might seem to prevent the semantics being used in any decision
procedure. But one could get around it by requiring, without loss of classical soundness, that the
language is ‘potentially infinite’ in the sense that it always contains at least one more sentence
letter than the formula whose status one wishes to decide (cf. the proof of soundness in section 6
below).

Downloaded from http://jigpal.oxfordjournals.org/ at University of Michigan on June 15, 2015


We may now formulate the completeness theorem: every classical tautology (in the language with
just →,⊥ as primitive connectives) is valid under Sandqvist’s semantics.
Remarks on the completeness theorem

(a) From his inferentialist perspective, Sandqvist sees this as a soundness theorem, stating the
soundness of the classical semantics wrt his inferential one. With classical logic as our ‘home
base’, we see it as a completeness theorem, reporting the completeness of Sandqvist’s seman-
tics wrt the classical one. This is a terminological contrast engendered by radically different
philosophical vantage points, but in no way affects the formal results.
(b) As noted by Sandqvist, the completeness theorem fails if ∨ is allowed as primitive with the
classical valuation B  ϕ ∨ψ iff B  ϕ or B  ψ, or the tighter valuation B  ϕ ∨ψ iff for all
C ⊇ B either C  ϕ or C  ψ, or the even more demanding one B  ϕ ∨ψ iff either for all C ⊇ B,
C  ϕ or for all C ⊇ B, C  ψ. Specifically, under each of those valuation procedures, excluded
middle p∨(p → ⊥) would fail at the empty basis, i.e. ∅  p∨(p → ⊥).
(c) However, the theorem continues to hold if ∨ is taken as primitive but given an evaluation
procedure that mimics the definition of ϕ ∨ψ as (ϕ → ⊥) → ψ, such as the following: B  ϕ ∨ψ
iff for all C ⊇ B, either (i) C  ψ or (ii) D  ϕ while p ∈
/ D(∅) for some D ⊇ C and sentence
letter p. Negation may likewise be taken as primitive, with an evaluation that similarly mimics
the definition of ¬ϕ as ϕ → ⊥: B  ¬ϕ iff for all C ⊇ B, if C  ϕ then C  p for every sentence
letter p. However, it seems unlikely that these rather complex recursion clauses can be further
simplified, or rendered more intuitive.
(d) Finally, we note that the theorem fails if ⊥ is given the classical-style valuation B  ⊥ for all
bases B while retaining Sandqvist’s evaluation of →. Specifically, double negation elimination
then fails at the empty basis, i.e. ∅  ((p → ⊥) → ⊥) → p. Verification: for every basis B there
is a basis C ⊇ B with C  p; we need only take C to be the set of all basic rules. Thus, under
the classical-style valuation of ⊥ combined with Sandqvist’s evaluation of →, there is no
basis B with B  p → ⊥. Hence, vacuously, ∅  (p → ⊥) → ⊥. Since also ∅  p we have ∅ 
((p → ⊥) → ⊥) → p.

3 Maxiconsistent bases
We do not hesitate to use full classical logic in our own reasoning, notably excluded middle for the
recursive definition of Bi+1 in the proof of Lemma 3.3. The first two lemmas are routine; the third
may be thought of as containing the ‘heart’ of the argument; the fourth reveals how the choice of
primitive connectives is critical.

[17:18 21/12/2013 jzt038.tex] Paper Size: a4 paper Job: JIGPAL Page: 149 147–154
150 Inferential semantics for classical logic
LEMMA 3.1
For bases B,C and formulae ϕ,ψ: (a) If B  ϕ and B ⊆ C then C  ϕ; (b) If B  ϕ and B  ϕ → ψ then
B  ψ; (c) If B  ⊥ then B  ϕ for all ϕ.
PROOF. Part (a) is the first lemma of Sandqvist [4] and is proven by an easy induction on ϕ. As a
corollary, it tells us that the validity of a formula ϕ in the semantics may equivalently be defined
as ∅  ϕ. Part (b) is immediate using the evaluation rule for →. For part (c), suppose B  ⊥. Then
by Sandqvist’s valuation of ⊥ we have B  p for every sentence letter p, and an easy induction on ϕ
gives us B  ϕ.
Remarks on Lemma 3.1

Downloaded from http://jigpal.oxfordjournals.org/ at University of Michigan on June 15, 2015


(a) Lemma 3.1 continues to hold if the language also has primitive ∧ with the classical-style
valuation B  ϕ ∧ψ iff B  ϕ and B  ψ.
(b) Kripke models also have the three properties of Lemma 3.1 (translating bases to states of the
model, and inclusion between bases to the relation between states). However, they also differ
in many respects. In particular, on the structural level, Sandqvist’s system uses only one model
(rather than a family of them) and this model has infinitely many points (all the uncountably
many bases), which are ordered by set inclusion (rather than by an arbitrary pre-order). On the
level of satisfaction, the status of sentence letters at a point is defined via the closure of the
empty set under a set of basic rules (rather than by a direct assignment), the falsum is given a
different valuation procedure and disjunction is not primitive.
We say that a basis B is inconsistent iff B  ⊥, otherwise B is consistent; and we call B maxiconsistent
iff it is maximal (under set inclusion) in the class of all consistent bases.
LEMMA 3.2
The following three conditions are equivalent for any basis B: (i) B is inconsistent, (ii) B  ϕ for
every formula ϕ, (iii) B  p for every sentence letter p.
PROOF. (i) implies (ii) by Lemma 3.1c, (ii) immediately implies (iii), and (iii) implies (i) by Sandqvist’s
evaluation of ⊥.
Remark on Lemma 3.2
(a) Of course, the implication from (iii) to (i) would fail under a classical-style evaluation of ⊥.
(b) If the language does not contain ⊥, leaving only →, then the equivalence of (ii) and (iii)
continues to hold for we can still infer (ii) from (iii) by an easy induction on ϕ.
The following lemma may be seen as the base case for its successor, which will pass from sentence
letters to arbitrary formulae.
LEMMA 3.3
For sentence letters p: if B  p there is a maxiconsistent basis C ⊇ B with C  p.
PROOF. First we show that if B  p there is a basis C that is maximal (under set inclusion) among
the bases X ⊇ B such that X  p. Then we show that it must be maxiconsistent, i.e. maximal (under
set inclusion) among the bases X ⊇ B such that X  ⊥.
For the first part, let σ1 , σ2 , … be an enumeration of all the countably many basic rules for the
language. Suppose B  p. Put C = ∪{Bi : i  0} where B0 = B, and Bi+1 = Bi ∪{σi } if Bi ∪{σi }  p else
Bi+1 = Bi . By induction, each Bi  p. Suppose for reductio that C  p. Then p ∈ C(∅) so p ∈ A(∅) for
some finite basis A ⊆ C. Since A is finite we have A ⊆ Bi for some i, so p ∈ Bi (∅) and thus Bi  p

[17:18 21/12/2013 jzt038.tex] Paper Size: a4 paper Job: JIGPAL Page: 150 147–154
Inferential semantics for classical logic 151

giving a contradiction. Thus C  p. Whenever D ⊃ C there is some σi ∈ D\C, so by construction


Bi ∪{σi }  p so D  p by Lemma 3.1a, giving maximality.
For the second part, since C  p Lemma 3.2 tells us that C is consistent; we need to show that
it is maximally so. To that end, we first claim that every basic rule p ⇒ q (with p as fixed, and
q freely chosen) is in C. Choose any sentence letter q. Since C  p we have p ∈ / C(∅), so C(∅) is
vacuously closed under the rule p ⇒ q and thus is closed under all rules in C + = C ∪{p ⇒ q}, so that
C + (∅) = C(∅) and p ∈ / C + (∅). Hence by the maximality of C among the bases not satisfying p,
+
we have C = C so p ⇒ q ∈ C as claimed. Now suppose D ⊃ C; we need to show that D  ⊥. By
the maximality of C among the bases not satisfying p, we have D  p, i.e. p ∈ D(∅), so since as
established p ⇒ q ∈ C ⊆ D for every sentence letter q we have every letter q ∈ D(∅), and thus, by
Sandqvist’s valuation of ⊥, D  ⊥ as desired.

Downloaded from http://jigpal.oxfordjournals.org/ at University of Michigan on June 15, 2015


Remarks on the proof of Lemma 3.3
(a) In the first part of the argument, the rule-compactness of C(·) is essential to get A ⊆ Bi for
some i.
(b) The construction C + = C ∪{p ⇒ q} in the second part is essentially the same as one used by
Sandqvist [4] in the proof of his Lemma 3.4 (page 217). For both his proof and ours, this
construction can be seen as the ‘heart’ of the argument.
(c) If the language does not contain ⊥ we can still carry through the proof of Lemma 3.3, redefining
inconsistency by either of the other two conditions noted to be equivalent in Lemma 3.2.
LEMMA 3.4
For any formula ϕ: if B  ϕ there is a maxiconsistent basis C ⊇ B with C  ϕ.
PROOF. We induce on ϕ. The base case is given by Lemma 3.3. In the induction step, there are two
cases to consider. For ϕ = ⊥: if B  ⊥ then by Sandqvist’s valuation of ⊥, there is some letter p with
B  p, and the base case again gives what we need. For ϕ = ψ → θ: if B  ψ → θ there is a B ⊇ B
with B  ψ but B  θ . From B  θ by the induction hypothesis there is a maxiconsistent basis C ⊇ B
such that C  θ. Since B ⊆ C and B  ψ, we have C  ψ by Lemma 3.1a. Putting together C  ψ and
C  θ gives us C  ψ → θ and we are done.
Remarks on the addition of connectives in Lemma 3.4
(a) If we add ∧ to the language as a primitive connective, evaluated in classical style by putting
B  ϕ ∧ψ iff B  ϕ and B  ψ, we can cover its case ϕ = ψ ∧θ in the induction step as follows:
Suppose B  ψ ∧θ . Then B  ψ or B  θ . Consider the former; the latter is similar. Since B  ψ,
the induction hypothesis tells us that there is a maxiconsistent C ⊇ B with C  ψ, so C  ψ ∧θ
and we are done.
(b) In contrast, if we add ∨ as primitive, evaluated in classical style by putting B  ψ ∨θ iff B  ψ
or B  θ , then the proof of Lemma 3.4 snags at the case ϕ = ψ ∨θ . For suppose B  ψ ∨θ . Then
B  ψ and B  θ . By the induction hypothesis there are maxiconsistent bases C, C  ⊇ B such that
C  ψ, C   θ , but we cannot conclude that either C  ψ ∨θ or C   ψ ∨θ .
The final lemma notes that in Sanqvist’s semantics with ⊥ and → primitive, maxiconsistent bases
behave essentially like classical models. Combining it with Lemma 3.4 will directly give Sandqvist’s
completeness theorem.
LEMMA 3.5
Every maxiconsistent basis C is classically well-behaved with respect to ⊥ and → in the sense that
C  ⊥ and for all formulae ϕ,ψ: C  ϕ → ψ iff C  ϕ or C  ψ.

[17:18 21/12/2013 jzt038.tex] Paper Size: a4 paper Job: JIGPAL Page: 151 147–154
152 Inferential semantics for classical logic

PROOF. We have C  ⊥ by the consistency of C. For → left to right: suppose C  ϕ → ψ and C  ϕ.


Then C  ψ by Lemma 3.1b as needed. Right to left: suppose C  ϕ → ψ. Then there is a C  ⊇ C
such that C   ϕ but C   ψ. Since C   ψ, Lemma 3.2 tells us that C  is consistent, so by the
maxiconsistency of C we have C  = C giving us C  ϕ but C  ψ as needed.

Remarks on good behaviour


(a) We needed the maxiconsistency of C to verify classical good behaviour of →, but only con-
sistency of C to get the same for ⊥.
(b) If we add ∧ as a primitive connective evaluated in classical style then, trivially, all bases (even
inconsistent ones) are classically well-behaved with respect to it.

Downloaded from http://jigpal.oxfordjournals.org/ at University of Michigan on June 15, 2015


Sandqvist’s Completeness Theorem: Every classical tautology (in the language with only → and ⊥
as primitive connectives) is valid in Sandqvist’s semantics.
PROOF. Suppose B  ϕ in Sandqvist’s semantics. We need to construct a classical valuation v such
that v(ϕ) = 0. By Lemma 3.4, there is a maxiconsistent basis C ⊇ B with C  ϕ. Let v be the
characteristic function of the set of formulae satisfied by basis C, i.e., put v(ψ) = 1 if C  ψ else
v(ψ) = 0, for all formulae ψ. Then v(ϕ) = 0, and by Lemma 3.5 it is a classical valuation for the
connectives → and ⊥.

Remarks on the argument as a whole


(a) As all lemmas continue to hold when the language also contains primitive ∧ with its classical-
style valuation, the completeness theorem also continues to hold in that context.
(b) However, as already noted, the proof blocks (at Lemma 3.4) when the language contains prim-
itive ∨ with a classical-style valuation.

4 Soundness of the semantics wrt classical logic


Sandqvist does not present a proof in his published paper of 2009 that every formula that is valid in
his semantics is a tautology, but refers the reader to Theorem 2.26 of the semi-published dissertation
of 2005. From his inferentialist perspective, he calls this the completeness of the classical semantics
wrt his semantics; from our classical perspective, we call it the soundness of the latter wrt the
former. The formal result is the same in each case, and we show that it also follows easily from our
Lemmas 3.3 and 3.5.
Soundness Theorem: Every formula (in the language with →, ⊥ as primitive connectives) that is
valid in Sandqvist’s semantics, is a classical tautology.

PROOF. Suppose ϕ is not a tautology; we want to show that B  ϕ for some basis B. Take a classical
valuation v with v(ϕ) = 0. Since ϕ contains only finitely many sentence letters and the language has
a countable supply of them (or, failing that, at least one more letter than those in ϕ, cf. the remarks
in section 2 on the supply of letters) we may assume wlog that v(q) = 0 for some letter q, which
we now fix. Put A = {∅ ⇒ p : v(p) = 1}∪{p ⇒ q : v(p) = 0}. Clearly q ∈ / {p : v(p) = 1} = A(∅), so A  q,
so by Lemma 3.3 there is a maxiconsistent B ⊇ A with B  q. From the latter and the definition of A
we have B(∅) = {p : v(p) = 1} = A(∅), so for every letter p, B  p iff v(p) = 1. By Lemma 3.5, since
B is maxiconsistent it is classically well-behaved wrt →, ⊥, and so we may conclude that B  ϕ iff
v(ϕ) = 1 and thus B  ϕ.

[17:18 21/12/2013 jzt038.tex] Paper Size: a4 paper Job: JIGPAL Page: 152 147–154
Inferential semantics for classical logic 153

5 Extension to first-order logic


The proofs of all lemmas, and the final verification for the completeness theorem, continue to go
through in the first-order context with primitive ∀, with sentence letters replaced by atomic formulae
and ∀ evaluated, as in Sandqvist [4], by the substitutional rule: B  ∀x(ϕ) iff B  ϕx:=t for all ground
terms t. Only routine additions are needed for Lemmas 3.4 and 3.5 and the final completeness
argument.
Specifically, for Lemma 3.4: in the case that ϕ = ∀x(ψ), suppose B  ∀x(ψ); then B  ψx:=t for
some term t, so by the induction hypothesis there is a maxiconsistent basis C ⊇ B with C  ψx:=t so
C  ∀x(ψ) and we are done. For lemma 3.5: under the evaluation ∀ behaves classically on all bases,
and so in particular on maxiconsistent ones. For the final check for the completeness theorem: as

Downloaded from http://jigpal.oxfordjournals.org/ at University of Michigan on June 15, 2015


usual, take the domain of the model to be the set of all ground terms of the language, and assign
values to predicates and function symbols in the natural syntactic manner to obtain a first-order
model v in which v(ψ) = 1 iff C  ψ, for all formulae ψ.

6 Extension to rules allowing empty right side


What would happen if we were to extend the notion of a basic rule by allowing its right-hand side
to be empty? This would need to be accompanied by a corresponding extension of the concept of
the closure of a set of sentence letters under a set of basic rules, as follows: B(K) is the intersection
of all sets L of sentence letters such that K ⊆ L and (a) whenever (Lf ,p) ∈ B with Lf ⊆ L then p ∈ L,
and (b) whenever (Lf ,∅) ∈ B with Lf ⊆ L then q ∈ L for every sentence letter q. Note that since ⊥
is not a sentence letter, B(K) never contains ⊥ (even when p ∈ K and p ⇒ ∅ ∈ B), and the operation
B(·) continues to enjoy all the nice properties noted in section 2.
Sandqvist’s completeness theorem continues to hold, as can be shown by a reduction argument. For
any basic rule ρ in the extended sense, we define a set ρ ∗ of basic rules in the original sense, as follows.
If the right side of ρ is non-empty, then ρ ∗ = {ρ}; if ρ = p1 ,...,pn ⇒ ∅ then ρ ∗ = {p1 ,...,pn ⇒ q : q a
sentence letter}. For any set B of basic rules in the extended sense, put B∗ = ∪{ρ ∗ : ρ ∈ B}. In words:
B∗ is obtained by replacing rules ρ = p1 ,...,pn ⇒ ∅ in B by the rules in ρ*. Clearly, B∗ is a basis in
the original sense of the term, and we have:

LEMMA 6.1
For any basis B in the extended sense, and any formula ϕ, B  ϕ iff B∗  ϕ.

Proof of Lemma 6.1


By induction on ϕ. For ϕ = p: It suffices to show that B(∅) = B∗ (∅), for which we need only observe
that, by definition, a set L of sentence letters is closed under a rule ρ = p1 ,...,pn ⇒ ∅ iff it is closed
under all the rules p1 ,...,pn ⇒ q with q a sentence letter. For ϕ = ⊥: We have B  ⊥ iff every sentence
letter is in B(∅) = B∗ (∅), iff B∗  ⊥. For ϕ = ψ → θ : Suppose B  ψ → θ . Then there is a C ⊇ B with
C  ψ and C  θ ; by the induction hypothesis C ∗  ψ and C ∗  θ , and clearly C ⊇ B implies C ∗ ⊇ B∗ so
B∗  ψ → θ . The converse is a little trickier. Suppose B∗  ψ → θ. Then there is a set D of basic rules
in the original sense such that B∗ ∪D  ψ and B∗ ∪D  θ . But since D is a basis in the original sense,
B∗ ∪D = (B∪D)∗ , so we have (B∪D)∗  ψ and (B∪D)∗  θ. Applying the induction hypothesis,
B∪D  ψ and B∪D  θ and thus we may conclude B  ψ → θ. 
Completeness theorem for generalized basic rules. Every classical tautology (in the language with
just →, ⊥ as primitive connectives) is valid in the above extended version of Sandqvist’s semantics.

[17:18 21/12/2013 jzt038.tex] Paper Size: a4 paper Job: JIGPAL Page: 153 147–154
154 Inferential semantics for classical logic

PROOF. If B  ϕ then B*  ϕ by Lemma 6.1, so by the completeness theorem in the original context
ϕ is not a tautology.

7 Philosophical perspectives
How far does Sandqvist’s semantics meet the conditions that inferentialists in the tradition of Prawitz
and Dummett would require for it to serve as a ‘justification’ of its outcome logic? This is a slippery
question to answer, for those requirements are not entirely clear. Indeed, it could be said that
Sandqvist’s completeness theorem reveals a situation that was quite unforeseen in that tradition, and
should force inferentialists to be more explicit about their acceptability criteria.

Downloaded from http://jigpal.oxfordjournals.org/ at University of Michigan on June 15, 2015


It might be suggested, as a criterion that would disqualify the semantics in the role of justification,
that every propositional connective that is expressed by a distinct word in informal mathematical
discourse should feature as primitive in the formal language of the inferential semantics, with an
evaluation rule that resonates with its intuitive understanding. Under this criterion, the primitives
should include not only ‘if’ and ‘and’ (which are covered by Sandqvist’s theorem as here extended)
but also ‘not’ and ‘or’ (which, as noted in section 2, are not covered unless given rather unintuitive
readings that mimic their definitions in the language of ⊥ and →).
However, the criterion is difficult to sustain. It is vague in so far as informal mathematical discourse
is expressed in somewhat regimented natural language, and different natural languages may convey
propositional connectives in different ways. Even for English, the status of multi-word connectives
like ‘neither … nor’, ‘exactly one of’, ‘only if’ and ‘if and only if’ (written as a single word ‘iff’
since the mid-twentieth century), is left unclear. Any specific resolution of this vagueness, obtained
by nominating a particular language (such as English) at a particular moment of its history (such
as today) and drawing a line among the various multi-word connectives, would appear to be rather
gratuitous. Formally, the inferentialist approach can lead to classical logic, or to various sub-classical
logics, according to the details of its formulation. Epistemologically, there does not appear to be
any convincing reason for disqualifying the former outcome.

Acknowledgements
Thanks to Lloyd Humberstone, Tor Sandqvist, Peter Schroeder-Heister and two anonymous referees
for valuable comments and suggestions.

References
[1] W. de Campos Sanz, T. Piecha and P. Schroeder-Heister. Constructive semantics, admissibility
of rules, and the validity of Peirce’s law. Logic Journal of the IGPL (to appear).
[2] D. Prawitz. Natural Deduction: A Proof-Theoretical Study. Almqvist & Wiksell, 1965,
reprinted 2006 by Dover Publications, Mineola, NY.
[3] T. Sandqvist. An Inferentialist Interpretation of Classical Logic. PhD Thesis, Uppsala Uppsala
Prints and Preprints in Philosophy, 2005.
[4] T. Sandqvist. Classical logic without bivalence. Analysis, 69, 211–218, 2009.

Received 30 June 2013

[17:18 21/12/2013 jzt038.tex] Paper Size: a4 paper Job: JIGPAL Page: 154 147–154

You might also like