Showing posts with label symbolic logic. Show all posts
Showing posts with label symbolic logic. Show all posts

Saturday, October 1, 2016

Beholder logic

I found this argument in the I Tyrant AD&D supplement and decided to try it out in symbolic form just because I could. Beholders are both seekers of knowledge and power and are also paranoids.

It is necessary to acquire knowledge at any cost.  I have knowledge others do not; some of these ignorant creatures are of the ideal breed; therefore, others of the ideal breed are out to get me.

Paraphrasing:

  1. For everyone and for all knowledge these don't have, then it is necessary to acquire it
  2. For everyone and for all knowledge  If I have it and it has to be acquired by them, then they are out to get me
  3. There's knowledge  I have and there are members of the ideal breed who don't have it

Therefore, there's a member of the ideal breed who's out to get me

Symbolizing:

Me=m
1
∀x(∀y[(Ky˄~Hxy)→Axy])
Premise
2
∀x(∀y(Ky˄Hmy˄Axy)→Gxm)
Premise (enthymeme)
3
∃y(Ky˄Hmy˄∃x(Bx˄~Hxy))
Premise
∴
∃x(Bx˄Gxm)
Conclusion


Alternate:

  1. For everyone and for all knowledge they don't have, then it is necessary to acquire it from anyone
  2. For everyone and for all knowledge, If it has to be acquired from anyone then they are out to get those
  3. There's knowledge  I have and there are members of the ideal breed who don't

Therefore, there's a member of the ideal breed who's out to get me


1
∀x(∀y[(Ky˄~Hxy)→∀zAxyz])
Premise
2
∀x(∀y∀z([Ky˄Axyz]→Gxz))
Premise (enthymeme)
3
∃y(Ky˄Hmy˄∃x(Bx˄~Hxy))
Premise
∴
∃x(Bx˄Gxm)
Conclusion

Both prove valid

Saturday, July 2, 2016

Fallacies in symbolic form

On this occasion I wanted to just give a few examples of argumentative fallacies and how they would look on symbolic form. Since symbols can cut through any haze, these can be used whenever possible to get clarity on what tricky arguments one is facing. Definitely not everything is here, but only some that have caught my eye. (Hope my symbols display well on your browser)

The Post Hoc

If something happens first (P) and something next (Q) one could conclude that
P→Q

4
P
From above
5
Q
From above
6
P→Q
?

Correct
4
P
From above
5
Q
From above
6
P˄Q
Conj 4,5

The most we can say is that something happens and that something else happens. 
P→Q might be True, but it cannot be logically derived from what we have here.

Appeal to Ignorance

You cannot prove you are not a communist, therefore you are one.

1
P→~Q
Premise (enthymeme. If you had proof, then you would not be...)
2
~P
P (You cannot prove…)
3
~~Q
From negating the antecedent 1,2
4
Q
Double negation 3 (Therefore you are a…)

Hmm...1 and 2 makes this one doubly wrong.

What about the ex silentio argument which is nonfallacious and looks similar?

1
Q→R
Premise  (If so-and-so was True, there would be evidence)
2
~R
Premise  (We've looked for the evidence and there's none)
3
~Q
MT 1,2 (So-and-so isn't True)[This one is valid]


Poisoning the well

These examples are taken from Madsen Pirie's How to Win Every Argument. Second lines are paraphrases. The first two seem to focus on the one making the proposal while the third on the proposal itself.

Everyone except an idiot knows…
Only an idiot doesn't know…

∀x(~Kx→Ix)
∀x(~Ix→Kx)
∀x(Ix ˅ Kx)
∀x~(~Ix ˄ ~Kx)
 ~∃x (~Ix ˄ ~Kx)


Only those who are inadequate now advocate…
Everyone except the inadequate do not advocate…

∀x(Ax→Ix)
∀x(~Ix→~Ax)
∀x(Ix ˅ ~Ax)
∀x~(~Ix ˄ Ax)
~∃x (~Ix ˄ Ax)


Choice in education is only a device…
The only thing that is choice in education is a device…

∀x(Cx→Dx)
∀x(~Dx→~Cx)
∀x(~Cx ˅ Dx)
∀x~(Cx ˄ ~Dx)
~∃x (Cx ˄ ~Dx)

Saturday, June 11, 2016

The Material Conditional and its truth values

... from that thing which is false what truth can come?

Ecclesiasticus 34:4

Time for something of more substance.
(Hope that the symbols display well on your browser)

The truth table for the material conditional is:

p
q
p ⊃  q
T
T
T
T
F
F
F
T
T
F
F
T

One can readily understand the first line and even the second line after a moment's reflection, but why does the conditional yield a True value for False antecedents? What follows are my attempts to make sense of it.

First try:


Eminent logicians agree that these are the values, so we are justified on relying on their opinion. This being logic however, we demand proof, so:

Second try:


"P then Q" is shorthand for ~P v Q or alternatively ~(P ^~Q). If one works out the truth table for these, the result is our starting table. However, still, at a gut level feels wrong: Out of Falseness, Truth?

Third try:


If one takes a step back, it makes sense that by starting with Falseness anything goes. One can draw anything from a False starting point, for instance "In a hole in the ground there lived a hobbit".  From here one can go anywhere, to Mt Doom, there, and back again. Now, more seriously:

Fourth try:


Going line by line one can see: first, having the condition of the antecedent met, we arrive at the consequent; that is the least we can expect of of any self-respecting conditional. Second, we don't want having the antecedent met, and not getting the consequent; the second line covers that. The last two lines makes clear that it doesn't care, it has no provision, when the antecedent is False.  The consequent may be True or it may be False, it makes no judgment other than let it pass. Parallely:

Fifth try:


So what is actually True for a conditional when its antecedent is False? Not the conclusion, but the expresion itself, the whole conditional. Having the antecedent False doesn't spring its magic , P then Q, leaving it aloof and unblemished ("You go do your nonsense; I'll keep to myself"). From another viewpoint:

Sixth try:


Russell said that for any two propositions p and q  either p⊃q or q⊃p. If you try a few in your head you'll see he's right in some sense. I think there's some controversy here.

Seventh try:


If you're still not convinced maybe you're right after all. Let's look at the alternatives and see if there' something better. Alternatives have been offered to fix the conditional and I'll look at just two of them here: giving the Truth table other values to better reflect natural language conditionals and getting into modal.

Other values:


There's an interesting article by Fulda0  which analyzes different possible truth values  for the conditional making it seven sets in all+ old material conditional. For my part implication #4 looked at first like the best candidate since its gives a False values for the False antecedents. It quickly becomes apparent that adopting it would bring even graver problems . Here's its Truth table:

p
q
 p ⊃  q (4)
T
T
T
T
F
F
F
T
F
F
F
F


For starters, the third line apparently makes the fallacies of denying the antecedent and affirming the consequent disappear. Then, we'd get some funky cause and effect relationships:

I aced Logic last year. The grape harvest in France shot up that same year. Then, If I ace Logic, the grape harvest will shoot up.

If next year the harvest looks chancy, I can picture France's grape growers in session drafting a letter earnestly asking me to hit the books and save their livelihood. I claim no such power.

What if we change back the third line?

p
q
 p ⊃ q (2)
T
T
T
T
F
F
F
T
T
F
F
F

If I shoot a man through the heart, then I'll kill him on the spot.

If I were to shoot him  and he dies for sure; so far so good.
If I shoot him through and he survives, that cannot be.
If I abstain from from shooting and he dies anyway; ok (maybe he had a medical condition)
It is not the case that I don't shoot him, and he doesn't die; that cannot be. Wait! Someone else could shoot him down for me. So having the fourth line all False is absurd.

Let's see if flipping the Truth value with the third helps:

p
q
 p ⊃ q (3)
T
T
T
T
F
F
F
T
F
F
F
T


The third line says that it can't be that he can continue living even when I abstain from shooting.  The table is tantamount to affirming that I have power of life and death over him and no one else. Once more, that's a bit much beyond me.


Going Modal:


Some philosophers were bothered at the time by the paradoxes material conditional (which are a different can of worms) or by how it could be strengthened. Notably, Lewis came up with the modern strict conditional and opened the field of modal logic. Are strict conditionals better?

According to Stanford1 paradoxes, now modal, popped up anyway and according to Girle2  "none of these [material] 'paradoxes' are problematic in and of themselves".

Also according to Girle3, to some, the material conditional appears to be no problem for math, so why bother?

Judging from Konyndyk4 what modal logicians regard as strict Truths,  it appears as if their field is very restricted especially since it doesn't include the usual suspect (ie physical necessity). Since logical Truths are included and material conditionals are tautologous, It appears to me that necessary Truths support the material rather than step away from it. I need someone shed some light on this point.

According to Barker5, a) Lewis overstated his case since Russell never drew anything in the Principia Mathematica from a False antecedent, ie he only drew Truth from Truth ; and b) Russell used the word 'implication' because he couldn't come up with a better term.



0 Fulda, J 2010, 'The Full Theory of Conditional Elements: Enumerating, Exemplifying, and Evaluating Each of the Eight Conditional Elements', Acta Analytica, 25, 4, pp. 459-477, Academic Search Complete, EBSCOhost, viewed 1 June 2015.
1 Sanford, D 1989, If P, then Q, Routledge, p73
2 Girle, R 2009, Modal Logics and Philosophy, Acumen, Durham, GBR. Available from: ProQuest ebrary. [2 June 2015] p89.
3 Girle, R 2003, Possible Worlds, Acumen, Durham, GBR. Available from: ProQuest ebrary. [2 June 2015] p33.
4 Konyndyk, K 1986, Introductory Modal Logic, University of Notre Dame Press; 1st edition
5 Barker, SF 2006, 'Lewis on Implication', Transactions Of The Charles S. Peirce Society, 42, 1, pp. 10-16, Academic Search Complete, EBSCOhost, viewed 1 June 2015.

Saturday, April 19, 2014

Logic Books: what to read and what to avoid

Not so long ago I tried to get head first into Aristotle's Organon starting backwards with the Topics  and, oh, it is so tough.

I blame his style for being so convoluted. It is still chock full of goodies, but much effort is needed to  unearth the gems. I decided that what was needed to get the info was either : a) sterling willpower and pen & paper; b) a guided tour;  or  c) the indirect route. I chose the latter by retaking symbolic logic, the same I should have learned from high school. Looking around, I found some books that I wanted to mention.

(What follows refers mostly to college level intro logic)

Sherlock's Logic


This is divided into two parts. The first one is a detective story involving a murder and the Templars. The story itself is good and a quick read. The second part introduces concepts of symbolic logic, one at a time, matching each with a chapter from the preceding story. Explanations are very condensed, which requires the reader to really exert herself to grasp each concept. In general, it is a survey of logic all right, but leaves out great many unexplored areas, has too few exercises (and those without answers), is unsystematic, steep,  and in the end, unsatisfying.

Introduction to Symbolic Logic by Karl J Smith


If you are on the run and need only a refresher, this is the book for you. Concepts are quickly explained and exercises for each chapter, with answers at the end of the book provided.

Schaum's Outlines: Logic 2nd by Nolt


Nolt's is just like Smith's but on steroids. You know the series. I came to this one for the solved exercises (which once you get to them are like meh), but stayed for the concise explanations which shine. This one covers the whole spectrum and a bit more. Inexpensive and readily available.

Logic: a Modular Approach by Facione & Scherer


Now we enter into those of textbook calibre. This one starts from naught and begins building from there, first with language and then with the usual logic programme. Each concept is clearly delineated and referenced to previous ones. As said the system is built upon itself through modules which correspond to subchapters in the usual sense. There are 46 modules in total. A fair amount of exercises are also to be found at the end of each module with answers to about half of the problems. Four cumulative tests are also included. This is an old book (the Disco font on the cover betrays its age), but still feels current and can still be purchased new. Compared to Copi's, this book is easier (yet not dumbed down) and its content more thoroughly explained. It unfortunately has also some long stretches of mind-numbing expositions. Its biggest faults however are that it only has a couple of modules on inductive reasoning and doesn't get into predicate calculus.

Introduction to Logic by Copi & Cohen


This appears to be the standard on logic textbooks. Notwithstanding, it is unnecessarily dense.   Answers to just 1 out of 5 exercises. I would skip this one altogether since, despite its many editions, it still needs moderate to heavy streamlining. It is also expensive.

Symbolic Logic by Copi


Tough and stingy with explanations, yet it is well worth a look to affirm and expand the basics.  Definitely not a first course. The hardest exercises are to be found here though this one too has too few solutions. Peter Suber and someone else on the net have solutions up to chapter 5.

Symbolic Logic: A First Course by Gary M. Hardegree


This one is the best on translation to symbolic (formalization) I've seen bar none. Goes up to polyadic predicate. It doesn't rush things and looks at every nook and cranny without losing the thread. Can take a while to make through the relevant chapters, but you end with the feeling of really knowing the stuff. Dozens of exercises are also provided with answers. Regrettably, identity is not covered.

Logic: Analysing and Appraising Arguments by Gensler


This one is an extensive attempt but is relatively light on explanations. Of all the books mentioned it is the only one which takes a look at modal logic in some length though. It has been superseded by a newer book by the same author. 

A Concise Introduction to Logic by Hurley


As with the first Copi, this is an inexplicably expensive book.  This one also carries about the same usual logic topics. Includes predicate logic calculus. Of all, this one felt best at expounding inductive logic and probably the best overall. Once more, answers to just a few exercises.

Logic by Paul Tomassi


In this one the author tries to make intro logic accessible to students who struggle with it once and for all. I'm not sure he succeeds in this, but the book might come in handy as a backup as it goes deeper than the rest on some philosophical aspects. As of late, the more I look at it, the more I like it.

Logic With Trees by Howson


A showcase of the Truth-Trees. The author takes a deep look into them which has the unfortunate effect of placing the book beyond the scope of first time delvers. Unless you really have to know about the inner-workings of Trees, you are better off with Tomassi's book or with Nolt's which look at their immediate, mechanical aspect in a more practical, concise manner. Howson's book could serve as a first course, but it wouldn't be my first choice, nor my second, nor my third, etc. Some solved exercises. Can be skipped safely.

First Course in Mathematical Logic by Suppes


This one too starts from the ground up, but shifts the focus from language to mathematical expressions. The author appears to be going somewhere with the focus on number expressions, but its the same logic from the other books with just different symbols. Worth a look if you're math oriented, have interest in logic in general or if it's the only one you can lay your hands on. If language or philosophy, however, is your interest, you might be better served  by one of the others. While strong on propositional,  for predicate it shrinks from Existential Instantiation/Elimination rules entirely. Note that set theory is also not covered. No answers to the exercises. Wah! Oh wait, Frederick Binford has those.

Lógica Clásica de Primer Orden by Falguera


The most dense of the ones here, idiosyncratic, out of print and in another language. All the same, it has great many substitution rules with proofs for predicate logic. Also has a couple hundreds of exercises with solutions on an accompanying booklet.

In closing, here's something completely different:

Logicomix by Doxiadis et al.


This is one is a graphic novel based on Bertrand Russell's search of Truth & certainty in logic. An easy read, places many of the concepts in their historical context and serves as a good appetizer for Gleick's The Information of which we'll talk next week.