Thursday 30 June 2011

The Logic Book, M. Bergmann, J. Moor, J. Nelson, McGraw Hill, 2004, 10.4E, 5(f), p. 557

We are asked to show that sentences in this pair are equivalent:
  1. (x)(y)[(Axy • Ayx) ⊃ Axx]
  2. (x)[(∃y)(Axy • Ayx) ⊃ Axx]
The proof:
  1. (x)(y)[(Axy • Ayx) ⊃ Axx]
  2. * (∃y)(Axy • Ayx) ......... ACP
  3. * Axm • Amx ......... 2 EI y/m
  4. * (y)[(Axy • Ayx) ⊃ Axx] ......... 1 UI x/x
  5. * (Axm • Amx) ⊃ Axx ......... 4 UI y/m
  6. * Axx ......... 3,5 MP
  7. (∃y)(Axy • Ayx) ⊃ Axx ......... 2-6 CP
  8. (x)[(∃y)(Axy • Ayx) ⊃ Axx] ......... 7 UG
and in reverse:
  1. (x)[(∃y)(Axy • Ayx) ⊃ Axx]
  2. * Axy • Ayx ......... ACP
  3. * (∃y)(Axy • Ayx) ......... 2 EG
  4. * (∃y)(Axy • Ayx) ⊃ Axx ......... 1 UI x/x
  5. * Axx ......... 3,4 MP
  6. (Axy • Ayx) ⊃ Axx ......... 2-5 CP
  7. (y)[(Axy • Ayx) ⊃Axx] ......... 6 UG
  8. (x)(y)[(Axy • Ayx) ⊃Axx] ......... 7 UG

Tuesday 21 June 2011

Symbolic Logic, D. Jacquette, Wadsworth, 2001, Chpt. 8, IV(19), p. 436

The argument is put like this:

"Either no multimillionaires are truly happy and satisfied with their lives, or some persons are dissatisfied with their lives. But everyone, at some level, is satisfied with his or her life. Therefore, all multimillionaires are dissatisfied with their lives."

The proof:
  1. (x){(Px • Mx) ⊃(y)[(Lyx ⊃ (¬ Hxy • ¬ Sxy)]} ∨(∃x)[Px • (∃y)(Lyx • ¬ Sxy)]
  2. (x)[Px ⊃(y)(Lyx ⊃Sxy)]
  3. ∴(x)[(Px • Mx) ⊃(y)(Lyx ⊃ ¬ Sxy)]
  4. (x){(Px • Mx) ⊃(y)[(Lyx ⊃ (¬ Hxy • ¬ Sxy)]} ¬ (x)[Px ⊃(y)(Lyx ⊃Sxy)] ......... 1 QC
  5. (x){(Px • Mx) ⊃(y)[(Lyx ⊃ (¬ Hxy • ¬ Sxy)]} ......... 2,4 DS
  6. * Px • Mx ......... ACP
  7. * * Lyx ......... ACP
  8. * * (Px • Mx) ⊃(y)[(Lyx ⊃ (¬ Hxy • ¬ Sxy)] ......... 5 UI x/x
  9. * * (y)[(Lyx ⊃ (¬ Hxy • ¬ Sxy)] ......... 6,8 MP
  10. * * Lyx ⊃ (¬ Hxy • ¬ Sxy) ......... 9 UI y/y
  11. * * ¬ Hxy • ¬ Sxy ......... 7,10 MP
  12. * * ¬ Sxy ......... 11 Simp.
  13. * Lyx ⊃ ¬ Sxy ......... 7-12 CP
  14. * (y)(Lyx ⊃ ¬ Sxy)] ......... 13 UG
  15. (Px • Mx) ⊃(y)(Lyx ⊃ ¬ Sxy)] ......... 6-14 CP
  16. (x){Px • Mx) ⊃(y)(Lyx ⊃ ¬ Sxy)]} ......... 15 UG

Thursday 16 June 2011

Tenses contra logic

The two terms are not contradictions but who has not thought as much at one time or another? Often it is not the question of tense that makes things difficult but the elusive aspect of English verbs: ‘I went by bus’ and ‘I always went by bus’. There is nothing in the verb ‘go’ that indicates regularity in the second sentence and a single event in the first.

Logic and tenses get in one another’s way on other levels too. The common sentence:

Either the payment has not been made or it has not been processed yet.

has the pattern of a disjunction: not P or not Q. If we reason with a disjunctive syllogism in mind, that is, by adding the premise: ‘The payment has been made,’ or P for short, then the conclusion is: therefore, it has not been processed yet. All’s fine. But if we add: ‘The payment has been processed,’ or Q for short, then we are forced to conclude that the payment has not been made. But how could the payment have been processed if it hadn’t been made?

With disjunction and conjunction, order does not matter in our logic and mathematics (there are logics where it does). ‘I took the money and went to the shop’ is the same as ‘I went to the shop and took the money’. Yet, just as frequency, order is a critical part of the English tense system.

Another oddity involves sentences such as:

My apple tree will produce a crop only if I spray it.

Here, ‘I spray’ is the necessary condition. In material conditionals, necessary conditions come in the main clause, so we can paraphrase the sentence as:

If my apple tree produces a crop, then I will spray it.

Immediately, we sense it is wrong. But it is not the logic that is wrong; rather the tenses got out of kilter in the shift. If we insist on this kind of paraphrase, then we should adjust the sentence to something like:

If my apple tree produces a crop, then I will have sprayed it.

Besides frequency and order, tenses convey causation – to which logic is largely indifferent. No one would think it odd to say:

If you drop it you’ll break it.

where clearly ‘dropping’ precedes ‘breaking’, and something of this order is captured in the tense sequence (present followed by future). But speculation as to the causes of the breaking might well turn up a sequence such as:

If you broke it, then you had dropped it. It is as simple as that.

The verbs have been reversed (and so the necessary condition is in the right place) but the order of events hasn’t.

Consider also: ‘You touch it, you bought it.’ Here the grammar misleads us, but the logic doesn’t. ‘Touching’ is a sufficient condition for ‘buying’, so it goes in the antecedent:

If you have touched, you have bought it.

Wednesday 15 June 2011

The Logic Book, M. Bergmann, J. Moor, J. Nelson, McGraw Hill, 2004, 10.4E, 1(j), p. 555

We are asked to derive the conclusion from the premises given.
  1. (y)(x)(Cxy ⊃ Qx)
  2. (y)(Qy ≡ Py)
  3. ∴(y)(x)(Cxy ⊃Px)
  4. (x)(Cxy ⊃ Qx) ......... 1 UI y/y
  5. Cxy ⊃ Qx ......... 4 UI x/x
  6. Qx ≡ Px ......... 2 UI y/x
  7. (Qx ⊃ Px) • (Px ⊃ Qx) ......... 6 BE
  8. Qx ⊃ Px ......... 7 Simp.
  9. Cxy ⊃Px ......... 5,8 HS
  10. (x)(Cxy ⊃Px) ......... 9 UG
  11. (y)(x)(Cxy ⊃Px) ......... 10 UG

Thursday 9 June 2011

The Logic Book, M. Bergmann, J. Moor, J. Nelson, McGraw Hill, 2004, 10.4E, 11(b), p. 559

We are asked to show that the following sentences are inconsistent: {(∃x)(Hx • Mxc); (x)(Lx ¬ Hx); (∃y)(Ly • Hy)}. In other words, we need to derive a contradiction. The thing to spot here is that we should start by instantiating the second of the existential statements (statement 3) and ignore the first. Otherwise we will not be able to instantiate statement 3 to the same constant as statement 2.
  1. (∃x)(Hx • Mxc)
  2. (x)(Lx ¬ Hx)
  3. (∃y)(Ly • Hy)
  4. La • Ha ......... 3 EI y/a
  5. La ......... 4 Simp.
  6. Ha ......... 4 Simp.
  7. La ¬ Ha ......... 2 UI x/a
  8. ¬ La ......... 6,7 MT
  9. ¬ La • La ......... 8,5 Conj.

Thursday 2 June 2011

Conclusion before proof

Proofs of arguments in logic and mathematics differ from problem solving (except where the latter is used loosely and covers both situations) in that we know the answer in advance. The goal is to show that, given the information we have, and the rules we know, we can actually reach the stated conclusion.

Starting with the conclusion is not cheating. We often work backwards, from the conclusion to the premises, as far as we can go, in order to spot any clues that might help us pick up the threads when we reason in the conventional but often ineffective way. In some contexts, our success depends to a large extent on whether we can actually articulate the conclusion, and then translate it into a language we can work with.

If, for example, the task is to prove that the composition of an odd and even function is always even, and we are actually able to put down on paper what has just been said, we have as good as proved the theorem. In this case, we’d be looking at proving that f(g(x)) = f(-g(x)). The proof consists of only a few steps and relies entirely on the definition of an odd and even function.

Deductive reasoning is in fact an infinitesimally small part of all the reasoning that we do, but it sets standards that are worth emulating despite one failed attempt after another. The professions pride themselves on reasoning for a living. The professionals I teach take umbrage at me saying they don’t reason deductively (I review and proofread legal matter as part of my job), but I don’t mean it as criticism. What I criticize, and which strangely they find uncontroversial at all, is that they reason the wrong way round. If they reason at all.

Typically, legal advice will be frontloaded with facts, evidence, legislation and legal interpretations, plus some assumptions, all of which is to serve as the premises. Then, writing and thinking start and proceed at the same time. This part is usually called ‘analysis’. That’s where the problem lies. The analysis should have been done before. The writing stage should be where the outcome of the analysis is summarized, not the analysis per se.

These reports, analyses, research papers and legal opinions often lack any sort of conclusion – they concentrate on the trip, not on the destination. The numerous ‘howevers’ and ‘neverthelesses’ are ways of telling the reader how much the writer is enjoying the trip. The writer says: let’s toss the facts ‘in the instant case’ and my rhetoric and let’s see what pattern they form when they fall on paper. When I finish the writing, I will know the conclusion. You wait patiently.

‘To the extent that the strategy you pursue may be contended (and constructed) to fit within the parameters of the law, success in prevailing against a challenge based upon alleged noncompliance will be enhanced, but this office is not in a position to predict the odds of such a successful outcome,’ is a legitimate conclusion, but if I have to wade through 5 pages of legal meanderings to read this, I want my money back.

Wednesday 1 June 2011

Symbolic Logic, D. Jacquette, Wadsworth, 2001, Chpt. 8, IV(18), p. 436

The argument:

"All giraffes are long-necked mammals of the African plains. Therefore, the tonsils of a giraffe are the tonsils of a long-necked mammal of the African plains."

The proof:
  1. (x)[Gx ⊃(Lx • Mx • Pxa)]
  2. (x)(y)[(Gy • Txy) ⊃ (∃z)(Lz • Mz • Pza • Txz)]
  3. * Gy • Txy ......... ACP
  4. * Gy ......... 3 Simp.
  5. * Gy ⊃(Ly • My • Pya) ......... 1 UI x/y
  6. * Ly • My • Pya ......... 4,5 MP
  7. * Txy ......... 3 Simp.
  8. * Ly • My • Pya • Txy ......... 6,7 Conj.
  9. * (∃z)(Lz • Mz • Pza • Txz) ......... 8 EG
  10. (Gy • Txy) ⊃ (∃z)(Lz • Mz • Pza • Txz) ......... 3-9 CP
  11. (y)[(Gy • Txy) ⊃ (∃z)(Lz • Mz • Pza • Txz)] ......... 10 UG
  12. (x)(y)[(Gy • Txy) ⊃ (∃z)(Lz • Mz • Pza • Txz)] ......... 11 UG