The Consequence Argument

The Consequence Argument

If determinism is true, then our acts are the consequences of the laws of nature and events in the remote past. But it is not up to us what went on before we were born, and neither is it up to us what the laws of nature are. Therefore, the consequences of those things (including our present acts) are not up to us. (van Inwagen 1983: 56)

• van Inwagen proposes three precisifications of this argument – perhaps it is rather ‘One Argument for Incompatibilism Done Three Ways’ (van Inwagen 1983: 56).
  • The SEP article has more detail (Vihvelin 2018, p. §5).
• We’ll focus on two versions, and take them in a different order: first, a modal argument van Inwagen labels ‘third’ (1983: 93–104), and then a non-modal argument that more closely follows the traditional terminology, which van Inwagen calls the ‘first’ argument (1983: 58–78).

Determinism Revisited

• van Inwagen (1983: §3.3) recasts the thesis of determinism:

  For every instant of time, there is a proposition that expresses the state of the world at that instant;
  If \(p\) and \(q\) are any propositions that express the state of the world at some instants, then the conjunction of \(p\) with the laws of nature entails \(q\). (van Inwagen 1983: 65)

• State of the world is a temporary, global, description of how everything is at an instant.

  • You might wonder how this fits with relativity theory, where there are no invariant facts about what is true ‘at an instant’. There is actually no problem: the relativistic laws are such that facts on a single simultaneity slice (a Cauchy surface) entail the contents of the whole Minkowski manifold (Earman 1986: 60). Things are trickier again in general relativity however (Earman 1986: 170–98).

• We might hesitate about entails; if so, how about: where \(\mathcal{L}\) is a proposition expressing the laws of nature, the propositions \(\mathcal{L}\), \(p\), and \(¬q\) cannot all be true together.

Could Have Rendered False

• There is a key notion that is important for connecting possible actions with determinism: having the ability to render a proposition false:

  to be able to render a proposition false is to be able to arrange or modify the concrete objects that constitute one’s environment – shoes, ships, bits of sealing wax – in a way sufficient for the falsity of that proposition.

  More precisely, we may define ‘\(s\) can render \(p\) false’ as follows:

  It is within \(s\)’s power to arrange or modify the concrete objects that constitute his environment in some way such that it is not possible in the broadly logical sense that he arrange or modify those objects in that way and the past have been exactly as it in fact was and \(p\) be true. (van Inwagen 1983: 67–68)

• This is the most general form of ability ascription; many outcomes I control indirectly, and this notion allows us to talk about all the things under my control, not only my immediate bodily movements.

Formalizing the Consequence Argument: Modal Version

Preliminaries

Let us suppose that there was a judge who had only to raise his right hand at a certain time, \(T\), to prevent the execution of a sentence of death upon a certain criminal,… Let us further suppose that the judge – call him ‘J’ – refrained from raising his hand at \(T\)…

I shall use ‘\(T_{0}\)’ to denote some arbitrarily chosen instant of time earlier than J’s birth, ‘\(P_{0}\)’ to denote a proposition that expresses the state of the world at \(T_{0}\), ‘\(P\)’ to denote a proposition that expresses the state of the world at \(T\), and \(\mathcal{L}\) to denote the conjunction into a single proposition of all the laws of nature. All these symbols are to be regarded as ‘rigid designators’. (van Inwagen 1983: 68–70)

• A rigid designator denotes the same thing in every possible situation – compare names like Antony (rigid) with descriptions like the lecturer for PHIL 2040 (non-rigid). So \(T_{0}\), \(P\), etc., are names of specific propositions.

Modal operators

• A modal operator like necessarily, or possibly, is a non-truth functional one-place sentential connective.

  • It’s a one-place sentential connective, like ‘\(¬\)’: if \(\phi\) is a sentence, so is necessarily \(\phi\), etc.
  • Semantically, it doesn’t express a truth function. If \(\mathcal{O}\) is a modal operator, the truth value of the constituent sentence \(p\) does not determine the truth of \(\mathcal{O}p\).

• We can demonstrate this by considering two sentences \(p\) and \(q\) which have the same truth-value, but where \(\mathcal{O}p\) and \(\mathcal{O}q\) differ in truth value.

• So let \(p\) be the falsehood \(2+2=5\), and \(q\) be the falsehood Antony is 50 years old. But:

  • Possibly, \(2+2=5\) (False);
  • Possibly, Antony is 50 years old (True – I hope).

• We can’t therefore capture the meaning of a modal operator using a truth table, unlike, e.g., \(¬\). But we can characterise it using inference rules. So, for example, using \(\square\) to represent necessarily, this rule is valid:

  (T)

  \(\square p \vdash p\).

The modal operator \(N\)

• Van Inwagen’s third (version of the) argument uses a distinctive modal operator \(N\). The intended interpretation of \(Np\) is this:

  \(p\) and no one has, or ever had, any choice about whether \(p\). (van Inwagen 1983: 93)

• If \(b\) represents the proposition the universe began with a Big Bang, then presumably \(b\) is true of our universe, and true that no one (God aside) ever had any choice about whether there was a Big Bang – we all came on the scene too late. So \(Nb\) is true.

• There are two rules which van Inwagen takes to govern the behaviour of \(N\):

  (\(\alpha\))

  \(\square p \vdash Np\). (Gloss: Whenever \(p\) is necessary, it’s also the case that \(p\) is true and no one ever had any choice about it. The converse does not hold: the Big Bang is contingent, even though no one had any choice about it, so \(Nb\) but \(¬\square b\). So \(N\) is a generalisation of strict necessity.)

  (\(\beta\))

  \(N(p \to q), Np \vdash Nq\). (Gloss: Whenever \(p\), and \(p\) implies \(q\), and no one had any choice about either of them, it’s also the case that \(q\) and no one had any choice about that either. This seems to hold because \(N\) is enough like necessarily, which does obey \(\square (p \to q), \square p \vdash \square q\).)

The ‘Third’ Formal Argument (van Inwagen 1983: 94–95)

1. \(\square ((P_{0} \wedge \mathcal{L}) \to P)\) (Follows from definition of determinism, because when \(\phi\) can’t be true together with \(¬\psi\), then if \(\phi\), \(\psi\) is necessary.)

2. \(\square (P_{0} \to (\mathcal{L} \to P))\) (Elementary logic of the material conditional ‘\(\to\)’, 1)

3. \(N(P_{0} \to (\mathcal{L} \to P))\) (2, rule \(\alpha\))

4. \(NP_{0}\) (premise)

5. \(N(\mathcal{L} \to P)\) (3, 4, rule \(\beta\))

6. \(N\mathcal{L}\) (premise)

7. \(NP\) (5, 6, rule \(\beta\))

Defending the Third Argument

• The conclusion (7) is incompatible with the judge having free choice over whether he raises his hand. Hence his action is unfree. The argument obviously can be generalised to any action, and hence supports incompatibilism.

• The options for the compatibilist, who wants to retain the possibility of both determinism and free action, are: reject premise (4); reject premise (6); reject rule (\(\alpha\)); or reject rule (\(\beta\)).

• The incompatibilist, like van Inwagen, has an easier response to the argument: accept the conclusion, or reject determinism (and hence premise 1). The latter is van Inwagen’s preferred route.

• Rule (\(\alpha\)) is unchallengable:

  I do not see how anyone could reject rule (\(\alpha\)). If (\(\alpha\)) is invalid then it could be that someone has a choice about what is necessarily true. (van Inwagen 1983: 96)

Defending (4) and (6)

I do not see how anyone could reject ‘\(NP_0\)’ or ‘\(NL\)’. My reasons are essentially those I gave in support of … premises [(12)] and [(13)] of the First Formal Argument. The proposition that \(P_{0}\) is a proposition about the remote past. We could, if we like, stipulate that it is a proposition about the distribution and momenta of atoms and other particles in the inchoate, presiderial nebulae. Therefore, surely, no one has any choice about whether \(P_{0}\). The proposition that \(\mathcal{L}\) is a proposition that ‘records’ the laws of nature. If it is a law of nature that angular momentum is conserved, then no one has any choice about whether angular momentum is conserved, and, more generally, since it is a law of nature that \(\mathcal{L}\), no one has any choice about whether \(\mathcal{L}\). (van Inwagen 1983: 96)

Defending (\(\beta\))

• Van Inwagen’s defence of (\(\beta\)) is to point out (i) its seeming self-evidence; (ii) its apparent correctness in other cases; and (iii) the reliance on it by others, particularly those who are involved in the biological determinism debate.

• Some compatibilists may deny (\(\beta\)); many did (e.g., Gallois 1977). Many such denials seem to beg the question (see van Inwagen 1983: 102–4).

  • If rejecting assumption because you accept the other assumptions and reject the conclusion is is question-begging, then any deductively valid argument is question-begging. A valid argument presents you with a choice: accept the conclusion, or reject the premises. Any valid argument will presuppose some premises which suffice to entail the conclusion, and so implicitly commit you to the falsehood of the conclusion. But this cannot be a dialectical flaw.

• So the charge of ‘begging the question’ must involve something more. The traditional fallacy of begging the question is: to accept the premises because you accept the conclusion. To reject rule (\(\beta\)) only because it leads to a conclusion you reject seems to be an instance of this informal fallacy.

• As van Inwagen accepts the premises because of their intuitive plausibility, he avoids the traditional charge.

Counterexamples to (\(\beta\))

• Some argue – by direct counterexample, not appeal to compatibilism – that (\(\beta\)) is invalid.

• For example, (\(\beta\)) entails this rule:

  (Agglomeration)

  \(Np, Nq \vdash N(p \wedge q)\).

  • This is an easy consequence of rules (\(\alpha\)) and (\(\beta\)) in the presence of this tautology of propositional logic: \((p \to (q \to (p \wedge q)))\). Since it’s a tautology, its necessary, so rule (\(\alpha\)) yields \(N(p \to (q \to (p \wedge q)))\); two applications of rule (\(\beta\)) given \(Np\) and \(Nq\) yield first \(N(q \to (p \wedge q))\) then \(N(p \wedge q)\).
    

• McKay and Johnson (1996) argue this rule has counterexamples; and thus (\(\beta\)) is problematic. Van Inwagen ends up agreeing: ‘Agglomeration is therefore invalid, and the invalidity of (\(\beta\)) follows from the invalidity of Agglomeration’ (van Inwagen 2000: 4).

Why is Agglomeration Invalid?

The following counterexample shows that the principle of agglomeration is invalid:

Suppose that I do not toss a coin, but could have.
\(p\) = the coin does not fall heads.
\(q\) = the coin does not fall tails.

Both premises of agglomeration are true, ‘\(Np\)’ and ‘\(Nq\)’; no one can choose to falsify \(p\) (no one can choose to make the coin fall heads) and no one can choose to falsify \(q\) (no one can choose to make the coin fall tails). The conclusion, ‘\(N (p \wedge q)\)’, is false, however. I could have chosen to make ‘\((p \wedge q)\)’ false by choosing to toss the coin, so I had a choice about whether ‘\((p \wedge q)\)’ is true. (McKay and Johnson 1996: 115)

Where to?

• Van Inwagen proposes to repair the argument by replacing \(N\) with another modal operator, one that does satisfy rule (\(\beta\)).
  • His candidate is this operator \(N^{*}\): ‘\(p\) and no one can, or ever could, do anything such that if she did it, \(p\) might be false’. (See the proposal to save \(\beta\) by redefining \(N\) (van Inwagen 2000: 8–10), and the rehearsal of the debate at (Vihvelin 2018: §5).)
• But this is all in fact a side issue. For the dispute over rule (\(\beta\)) leaves the other formalizations of the Consequence Argument untouched; they are not modal arguments, and hence do not involve this rule, and van Inwagen could just say ‘I now reject (\(\beta\)), but still endorse the other arguments’.
• So the real debate concerns the traditional version of the consequence argument, and the most substantial objection is not to the superficial formalization of the argument, but a deeper reason to doubt all versions of the Consequence Argument: the argument in Lewis (1981) against premise (6) and its analogs (see also Vihvelin 2013: 162–66).

Formalizing the Consequence Argument: ‘Traditional’ Version

‘First’ Formal Argument (van Inwagen 1983: 70)

8. If determinism is true, then the conjunction of \(P_{0}\) and \(\mathcal{L}\) entails [or necessitates] \(P\). (premise, definition of determinism)

9. It is not possible that: J could have raised his hand at \(T\) and \(P\) be true. (premise, definition of \(P\))

10. If J could have raised his hand at \(T\), J could have rendered \(P\) false. (From 9, logic)

11. If J could have rendered \(P\) false, and if the conjunction of \(P_{0}\) and \(\mathcal{L}\) entails \(P\), then J could have rendered the conjunction of \(P_{0}\) and \(\mathcal{L}\) false. (premise)

12. If J could have rendered the conjunction of \(P_{0}\) and \(\mathcal{L}\) false, then J could have rendered \(\mathcal{L}\) false. (premise)

13. J could not have rendered \(\mathcal{L}\) false. (premise)

14. If determinism is true, then if J could have raised his hand at \(T\), J could have rendered \(\mathcal{L}\) false. (8, 10, 11, 12)

15. If determinism is true, J could not have raised his hand at \(T\). (14, 13)

Defending the Argument: premises (8), (9), (11)

• This argument is valid; van Inwagen contends it uses no modal operators, but arguably it uses counterfactual conditionals at (10), (11) and (12). But the argument is valid in orthodox counterfactual logics.
• Premise (8) and (9) are obviously true, from the definitions of determinism, and the propositions involved.
  • Van Inwagen does note that the ability involved in (9) is, using his disambiguating notation, ‘J could have (raised his hand at \(T\))’; (10) follows given the principle of counterfactual logic that \(\square \neg(\phi \wedge \psi) \vdash \phi > \neg\psi\).

premise [(11)] is an instance … of the following… principle:

If \(s\) can render \(r\) false, and if \(q\) entails \(r\), then \(s\) can render \(q\) false.…

This principle is a trivial truth. For if \(q\) entails \(r\), the denial of \(r\) entails the denial of \(q\). Thus anything [including any arrangement of objects \(s\) can produce] sufficient in the broadly logical sense for the falsity of \(r\) is also sufficient for the falsity of \(q\). (van Inwagen 1983: 72)

Defending the Argument: premise (12)

• Premise (12) is an instance of:

  If \(q\) is a true proposition that concerns only states of affairs that obtained before \(s\)’s birth, and if \(s\) can render the conjunction of \(q\) and \(r\) false, then \(s\) can render \(r\) false. (van Inwagen 1983: 72)

• This seems right: if J had raised his hand, then the distant past would (still) have been the same; and so if J had raised his hand, its the laws that would have been different – given that one or the other of them would have to be different.

  • Here we rely on the principle that if it had been that \(p\), then it would have been that \(q\) entails if \(p\) then \(q\).

• Van Inwagen supports this by considering a case he regards as analogous: if I am to render false the conjunction of a historical and a non-historical proposition, it must be by rendering false the non-historical proposition.

Defending the Argument: premise (13)

• This seems to follow immediately from whatever a law of nature is: given that the laws of nature are – at the very least – non-accidental exceptionlessly true regularities, I cannot break a law, or through any action of mine, cause a law to be broken:

  if human beings can (have it within their power to) conduct an experiment or construct a device that would falsify a certain proposition, then that proposition is not a law of nature. … the laws of nature impose limits on our abilities: they are partly determinative of what it is possible for us to do. (van Inwagen 1983: 62)

Lewis on the Consequence Argument

Compatibilism and Law-Breaking

• Any compatibilist, who thinks that J could have raised their hand, is thereby logically committed to the antecedent of the conditional premise (12) – for J’s raising his hand would ensure that at least one of \(P_{0}\) and \(\mathcal{L}\) is false.

• Since this conditional obeys modus ponens, compatibilism appears to entails that J could falsify a law of nature:

  [As a compatibilist] who accepts the requisite auxiliary premises and principle of counterfactual logic, I am committed to the consequence that if I had done what I was able to do – raise my hand – then some law would have been broken. (Lewis 1981: 114)

Refining the Conclusion

“That is to say,” my opponent paraphrases, “you claim to be able to break the very laws of nature. And with so little effort! A marvellous power indeed! Can you also bend spoons?”

Distinguo. My opponent’s paraphrase is not quite right. He has replaced the weak thesis that I accept with a stronger thesis that I join him in rejecting. The strong thesis is utterly incredible, but it is no part of soft determinism. The weak thesis is controversial, to be sure, but a soft determinist should not mind being committed to it. The two theses are as follows.

(Weak Thesis)

I am able to do something such that, if I did it, a law would be broken.

(Strong Thesis)

I am able to break a law.

Rendering the Laws of Nature False

• What Lewis points out here is a potential ambiguity in van Inwagen’s crucial phrase, could have rendered false.
• The thought is then that the Consequence Argument equivocates: the consequent of premise (12) uses the phrase in one way, the weak way; but the argument that van Inwagen offers for premise (13) – that J is not able to render a law false – rests on the implausibility of the strong thesis.
• But we now have to grasp the distinction Lewis is drawing – and it’s a subtle one.
• The key observation is that the weak thesis involves a counterfactual conditional, related to the conditional in premise (12), and the correct understanding of this conditional will help us see why (12) might be true and (13) false.

Counterfactuals

A counterfactual, ‘If it were that \(A\), then it would be that \(C\)’ is (non-vacuously) true iff and only if some (accessible) world where both \(A\) and \(C\) are true is more similar to our actual world, overall, than is any world where \(A\) is true but \(C\) is false. (Lewis 1979: 465)

16. If kangaroos had no tails, they would fall over.

• (16) is true; it is true, according to Lewis (1973) and Stalnaker (1968), because
  1. There is a possible situation very much like actuality except that kangaroos have recently lost their tails due to misadventure or malign action;
  2. Any possibility that differs from actuality both in removing the tails of kangaroos and then attempting to restore the kangaroos’ prior stable mass distribution would be yet more dissimilar from actuality; so
  3. All the most similar situations to actuality involve unbalanced kangaroos, which of course are prone to falling over.

Evaluating Counterfactuals: a heuristic discerned in our practice

(C1)

It is of the first importance to avoid big, widespread, diverse violations of law.

(C2)

It is of the second importance to maximize the spatio-temporal region throughout which perfect match of particular fact prevails.

(C3)

It is of the third importance to avoid even small, localized, simple violations of law.

(C4)

It is of little or no importance to secure approximate similarity of particular fact, even in matters that concern us greatly. (Lewis 1979: 472)

• These criteria are supposed to be descriptive of our actual usage of counterfactuals, not some a priori philosophical prescription.
  • When we consider had I caught the ball, the window wouldn’t have broken, we envisage a constant history with just enough deviation to secure my catching the ball, and then the laws of nature resume thereafter. We don’t typically accept had I caught the ball, the whole past would have been different.

Divergence Miracles

• Lewis’ proposal is this: the most similar worlds to actuality are those where, to achieve the truth of the antecedent \(A\) of a counterfactual \(A > C\), we envisage a small divergence miracle – a small deviation from the actual laws – just before the time of \(A\), and which suffices to make \(A\) true.
• The laws of these similar worlds are not the laws of the actual world, obviously. But if criteria (C1) and (C2) can be satisfied, such a world can still be amongst the most similar – more similar than any world where the laws are exactly the same as the actual world.
  • For those worlds secure complete match of laws – satisfying (C3) at the expense of the higher-ranked (C2).
• Obviously this ranking fits well with Humean approaches to laws of nature as supervening upon occurrent fact (Lewis 1994: 478–80), but doesn’t rely on it.

Return to Law-Breaking

Had I raised my hand, a law would have been broken beforehand. The course of events would have diverged from the actual course of events a little while before I raised my hand, and, at the point of divergence there would have been a law-breaking event—a divergence miracle…. But this divergence miracle would not have been caused by my raising my hand. If anything, the causation would have been the other way around.… To accommodate my hypothetical raising of my hand while holding fixed all that can and should be held fixed, it is necessary to suppose one divergence miracle, gratuitous to suppose any further law-breaking.

Thus I insist that I was able to raise my hand, and I acknowledge that a law would have been broken had I done so, but I deny that I am therefore able to break a law. (Lewis 1981: 116–17)

Commentary on the Foregoing

• Lewis argues that to evaluate any conditional of the form If I had raised my hand, then \(p\), we need to consider the closest possible situations in which I raise my hand.

  • By the criteria, scenarios in which the laws are slightly different from actuality – just enough to permit me to raise my hand consistently with the actual initial conditions – are closer (more similar) than any scenarios in which I raise my hand and where the actual laws of nature remain in effect.

• For someone to break a law

  1. There must be some action \(A\) that, given the background facts, is inconsistent with some actual law of nature \(\mathcal{L}\), and
  2. The person performs \(A\) in a situation in which \(\mathcal{L}\) is a law.

• Since the laws are at least exceptionless regularities, it is not possible to break a law in this sense; no one can satisfy clause (ii), breaking a law while it remains a law.

• The weak thesis doesn’t require clause (ii), but rather, as the discussion of counterfactuals just showed, only this weaker claim:

  3. the person performs \(A\) in a situation \(w\) where, though \(\mathcal{L}\) is an actual law, \(\mathcal{L}\) is not a law in \(w\).

The Consequence Argument Revisited

Van Inwagen’s argument runs as follows, near enough. (I recast it as a reductio against the instance of soft determinism that I feign to uphold.) I did not raise my hand; suppose for reductio that I could have raised my hand, although determinism is true. Then it follows, given four premises that I cannot question, that I could have rendered false the conjunction \(H\mathcal{L}\) of a certain historical proposition \(H\) about the state of the world before my birth and a certain law proposition \(\mathcal{L}\). If so, then I could have rendered \(\mathcal{L}\) false. (Premise [12].) But I could not have rendered \(\mathcal{L}\) false. (Premise [13].) This refutes our supposition. (Lewis 1981: 118–19)

• Note: Lewis explicitly treats van Inwagen’s could have rendered false construction counterfactually.

Counterfactuals and Law-Breaking

• Lewis invites us to consider a counterfactual: had I raised my hand, the conjunction \(H\mathcal{L}\) would have been false.

• Accordingly, one of these counterfactuals must be true:

  Different Past

  If I had raised my hand, the remote past would have been different.

  Different Laws

  If I had raised my hand, the laws would have been different. (Vihvelin 2013: 164).

• On Lewis’ view, it is more important to keep the past the same than to keep the exact laws; so he rejects Different Past, but accepts:

  Slightly Different Laws

  If I had raised my hand, the laws would have been ever so slightly different in a way that permitted the occurrence of a lawful divergence from actual history shortly before the time of my decision (Vihvelin 2013: 165).

Lewis on the Consequence Argument: the equivocation revealed

The Weak Thesis … is the thesis that I could have rendered a law false in the weak sense. The Strong Thesis, which I reject, is the thesis that I could have rendered a law false in the strong sense.

The first part of van Inwagen’s argument succeeds whichever sense we take. If I could have raised my hand despite the fact that determinism is true and I did not raise it, then indeed it is true both in the weak sense and in the strong sense that I could have rendered false the conjunction \(H\mathcal{L}\) of history and law. But I could have rendered false the law proposition \(\mathcal{L}\) in the weak sense, though I could not have rendered \(\mathcal{L}\) false in the strong sense. So if we take the weak sense throughout the argument, then I deny Premise [13]. If instead we take the strong sense, then I deny Premise [12]. (Lewis 1981: 120)

The Dialectic

• There has been considerable debate over whether Lewis’ criteria really do manage to successfully capture the intuitive truth-conditions of counterfactuals (Fine 1975; Wasserman 2006).

• Vihvelin argues that Lewis’ response is actually insensitive to this issue – that Lewis’ conclusion is that we are able to do law-violating things, but these are abilities

  which we would exercise only if the past (and/or the laws) had been different in the appropriate ways. And while this may sound odd, it is no more incredible than the claim that the successful exercise of our abilities depends, not only on us, but also on the co-operation of factors outside our control. (Vihvelin 2013: 165–66)

Contextualism Revisited

Lewis on Fatalism, Again

Fatalists – the best of them – are philosophers who take facts we count as irrelevant in saying what someone can do, disguise them somehow as facts of a different sort that we count as relevant, and thereby argue that we can do less than we think – indeed, that there is nothing at all that we don’t do but can. I am not going to vote Republican next fall. The fatalist argues that, strange to say, I not only won’t but can’t: for my voting Republican is not compossible with the fact that it was true already in the year 1548 that I was not going to vote Republican 428 years later. My rejoinder is that this is a fact, sure enough; however, it is an irrelevant fact about the future masquerading as a relevant fact about the past, and so should be left out of account in saying what, in any ordinary sense, I can do. (Lewis 1976: 151)

Does this Apply to Incompatibilism?

• Lewis’ suggestion is that some facts that trivialize can claims should nevertheless be ignored in ascribing abilities.
  • In the fatalist case, facts about the outcomes of my actions are facts that should be ignored in thinking about abilities.
• The semantics of can permit this, as can \(\phi\) is true iff \(\phi\) is compatible with the contextually relevant background facts (Kratzer 1977) – we just need to note that, in most contexts, the facts about my future actions aren’t conversationally relevant.
  • In time travel cases, of course, Lewis thinks the typical relevance of facts about my personal past, and the typical irrelevance of facts about the future, clash and produce conflicting intuitions.
• The determinist takes facts about the distant past, and attempts to use the laws of nature to make those facts relevant to our present actions – arguing likewise that we can do less than we think.
• Can we use the role of can in ability ascriptions, along the lines of the contextualist theory of free will to resist the determinist, just as we resist the fatalist?

Incompatibilism and Context

(IC)

If determinism is true, the only outcomes which can happen, given the past, are those which do happen.

• Incompatibilists who use (IC) to express their view are not trying to make a linguistic claim. But if the standard view is right about can, the sentence (IC) has a subtle context-sensitivity.

• To defend (IC), despite the context-sensitivity it involves, would involve arguing that facts about the complete distant past and the laws are always contextually salient – they can never be properly ignored (Lewis 1996: 554).

• A sort of compatibilism can be defended if there are contexts in which the proposition expressed by (IC) is false, in which some of the past or the laws can be properly ignored:

  Such contexts exist, and … it would certainly be premature to reject compatibilism on the basis that there exist other contexts where (IC) is true. My own view is that compatibilism is the correct attitude in ordinary contexts, largely because ordinary unreflective use is clearly compatibilist, and there is no evidence that uses of can … invariably force the whole past and the laws to be relevant in a context. (Eagle 2011: 286)

Contextualist Compatibilism: A Concessive Response

• This is a concessive response: it concedes that, when the past and laws are relevant, the thesis of incompatibilism expresses a truth.
• But it denies that incompatibilism (IC) always expresses a truth. In those contexts where it does not, some things can happen which will not is true, and so (disquoting) some things can happen which will not, i.e., given determinism, some things can happen that the past determines will not happen.
• This response is viable if, despite the truths about determinism and the past, they can sometimes, at least in part, be properly ignored.
• Perhaps someone might argue that the truth of determinism should always be relevant, and can never properly be ignored – but why?
  • We do not concede to the traditional fatalist that future facts should always be relevant, despite their obvious pertinence to the outcome in question.
  • And on at least some (‘Humean’) views, facts about the laws of nature are disguised facts about the future (Barnes and Cameron 2009: 300–301; Eagle 2019: §6): whether \(\mathcal{L}\) is a law depends in part on whether future things act in accordance with \(\mathcal{L}\).

The Consequence Argument Again

• Let’s focus on premise (13): J could not have rendered \(\mathcal{L}\) false.
• As long as there are contexts where we don’t hold the laws fixed in evaluating what someone can do, then (13) will not express a truth in every context.
• Lewis’ principles governing similarity will arguably yield falsifying contexts: in a context where what we hold fixed is governed by (C1)–(C4), the laws are not held fixed.
  • Since counterfactuals themselves are notoriously context-sensitive (Lewis 1973; van Fraassen 1981), it is very plausible that Lewis’ criteria do give the correct similarity ordering on some occasions.
• Interestingly, these contexts make the argument unsound. But they are not contexts in which the argument could be unsoundly given – indeed, contextualist compatibilism holds out the prospect that the Consequence Argument is always valid and sound in contexts where it is uttered.
  • Given the way context works, if those facts are not ignored, they are not properly ignored (Lewis 1996: 559). So whenever determinism is mentioned, as it is in an utterance of (IC), it is salient. This might explain the philosophical appeal of incompatibilism, despite the fact that we never appear to acknowledge it in ordinary life.

References