Wednesday, April 11, 2012

More Geometrico

Philosophy, linguistics, mathematics and computer science are just stupid names for the bureaucrats. As Árpád Szabó showed in his works (e.g. in The Beginnings of Greek Mathematics) the logical investigations of the Eleatics school were a necessary precondition of the birth of axiomatic mathematics. Euclid's Elements is a real gem (Smullyan writes in superlatives about the book in his autobiography, and according to him, it is still the best introduction into geometry and mathematical thinking) and it was unquestionable till the 19th century.

In 1637 Descartes published the Discours de la méthode with a less known appendix titled La Géomtérie in which he showed his calcul géométrique, today this field is called analytic geometry or coordinate geometry, sometimes Cartesian geometry. The axiomatic method of geometry deeply influenced Descartes, and I guess Descartes mathematics was influenced by the Cartesianism too. Spinoza continued the Cartesian line and built up his ethics more geometrico (in the geometrical manner).

The foundational crisis of mathematics was preceded by "mini-crisis" in geometry. János Bolyai and Nikolai Lobachevsky "discovered" the non-euclidian geometry almost at the same time. Bolyai wrote in a letter to his father "by denying the axiom of parallels i made a wonderful new world from nothing". The philosopher of mathematics and mathematician Imre Toth calls this process of "creating negation", and he thinks this is the differencia specifica of mankind; we can make new worlds by the power of thinking (if you want to read non-analytic philosophy of maths you should read Toth's books, my personal favorites Platon et l'irrationnel mathématique and Liberté et vérité: Pensée mathématique et spéculation philosophique. If you want to read a postmodern quasi-literary maths book, I recommend you Toth's Palimpseste: Propos avant un triangle ).

As Wadler says in his Proofs Are Programs "For my money, Gentzen’s natural deduction and Church’s lambda calculus are on a par with Einstein’s relativity and Dirac’s quantum physics for elegance and insight." The foundational crisis brought us Frege's works. The logicisim was a reaction to psychologism and neo-kantianism, not "just" yet another mathematical theory. The early 20th century was about finding new ways. The Vienna Circle, the Prague Circle and the Polish School of Mathematics along the flourishing Geman and Hungarian mathematics were exploring the new more geometrico. These schools left the old continent and new ones emerged. For me, generative linguistics is the descendant of the Vienna and Prague circles, highly influenced by the Polish school.

Ethics and political philosophy went back to its roots when Rawls' A Theory of Justice was published.Rawls' linguistic analogy is just one member of the bestiary of formal methods in ethics and political philosophy. These days, you can find "hard core techniques" if you search for social software and dynamic epistemic logic.

During the the second half of the last century new sciences born like computer science, artificial intelligence, cognitive science. These brand new sciences are trying to answer old philosophical questions more geometrico but can we draw a line between science and philosophy?

Monday, April 9, 2012

Preparing for the summer

This summer is about learning. I'm very happy and fortunate to work at Weblib because they sponsor my dream - i can attend summer schools :D I'm going to attend a few events too - as part of my transition from an armchair thinker into an "applied researcher".

For me, the summer starts in May with the first Hungarian Natural Language Processing Meetup. I'm the chief organizer of the meetup and I do hope we'll have fun on 10 May! I'm going to talk about computational linguistics and Clojure at the EuroClojure 2012 conference late May. In mid June I'm visiting the Riviera Scala Clojure Group (thank to @anrizal for inviting me) and I'll give a talk (and I hope I'll make new friends too).

I've applied for the LXMLS (Lisbon Machine Learning Summer School). I hope I can attend it (this summer school seems to be very selective), because I've never learned ml in school and it would be a great opportunity to consolidate my knowledge.

I'm going to attend ESSLLI 2012 this summer and I'm preparing myself for the event. I think two weeks of intensive learning must be tough, so I'd like to read as much as possible in advance. First, I chose non-overlapping courses. During the next few weeks, I try to get a first impression about the topics and I'll eliminate at least one course from each week's list.

Week 1
  • Probabilistic Models in the Study of Language, Roger Lev - textbook in progress by Lev
  • Epistemic Game Theory
  • Purely Functional Algorithm Specification
  • Distributional semantics for linguists
Week 2
  • Models of Computation
  • Circularity
  • Annotation Mining with R
  • Social Laws: Logic and Games

Sunday, April 8, 2012

Set Theory 1.

These days, I'm reading about set theory because I realized I know almost nothing about this topic. Our intern asked me to help him in preparing for his MA in Logic and Theories of Science so I we made a kind of curriculum. Given the short time till the entrance exam, we are focusing on the very basics of logic. We are using the Vol. 1 of Gamut and Part I. of Partee et al. Although I really enjoy learning with our intern and I love discussing some questions with him, I'd like to deepen my knowledge of the subject.

The first step, as I mentioned above, is part one of the classic Mathematical Methods in Linguistics. I guess a mathematician would call it "set theory for dummies", but sometimes one needs a lightweight intro to see the 'big picture'. I considered myself very good at set theory but I didn't regret rereading this material. The next natural step was Naive Set Theory by Halmos - the book was on my shelf for years, as my adviser recommended it but I completely forgot it.

You can find pretty good articles on the Stanford Encyclopedia of Philosophy, here are a few:
Speaking about set theory one can't ignore Quine's Set Theory and Its Logic. Frege's Foundations of Arithmetic is a real gem, and although one can enjoy it with only a little background in set theory and/or logic, you will find it more interesting and deep with more knowledge under your belt. If you are into philosophy, I strongly recommend you Potter's Set Theory and Its Philosophy.

I ordered two books to continue my adventures in set theory, but I haven't received them yet. An old chap recommended The Philosophy of Set Theory: An Historical Introduction to Cantor's Paradise and Suppes' Axiomatic Set Theory as good intermediate books and I hope I can start reading them soon.

Thursday, February 16, 2012

A History of Haskell

I got interested in the history of Haskell again. In my previous post - The Roots of Haskell - I cited very theoretical papers on lambda calculus and I missed something like Graham's The Roots of Lisp. I recently came across "A History of Haskell: being lazy with class" by Paul Hudak, John Huges, Simon Peyton Jones and Philip Wadler, and thanks to a paywall that prevented me from downloading it I had to google for other sources (the link above points to a free copy) so I find a video lecture with the same title (the file is 630MB!). Hope you'll enjoy it.

 

P.S.: I'm trying to post more often in the future :D

Friday, November 25, 2011

A Philosopher's Road to Haskell

In the last 8 months, I was guiding a friend of mine (Zsolt Varga) through Haskell. Zsolt is an excellent logician and philosopher, he finished his PhD this year and started two part-time jobs, so he's a part-time web developer, and  a part-time lecturer in philosophy and he thinks these two seemingly different jobs are very close to each other. My dear friend asked me if I'm still into functional languages because he wants to learn Haskell. We talked a lot about where he should start and etc. and at the end I realized this would be a great opportunity to consolidate my knowledge too. We held our online meetings once or twice a week in the mornings and I guess our experiences might be of your interest.

Wednesday, November 16, 2011

Why Clojure lx?

This post is cross-posted on clojurelx, a new project blog

The NLTK is a natural choice for students of linguistics and computer science. It has matured into a stable project, its users are very active, and it is now used outside of academia. Those who are into functional programming can use the Scheme Natural Language Toolkit, or learn from the Natural Language Processing for the Working Programmer, and those who needs the JVM can turn to ScalaNLP. So why brother with Clojure?

Sunday, September 25, 2011

[Draft] One Culture: Chomsky, Norvig, Turing, etc.

[This post is a draft - it is incomplete, inconsistent, and I jump from one topic to the other without any logic. Please help me to make it better and even finalize it, so leave comments.]

Earlier this year, I had a debate on Twitter, right after the TR article and Norvig's essay, and Jeremy Kahn wrote a post about this correspondence. In the comments, I wanted to emphasis that Chomsky's position in linguistics is very 'special' as he's a successor of a very different school, the Vienna Circle and the two "paradigms" are really one. Interpreting the history of science is not an exact science by any means, but I'd like to sketch a story that shows the two seemingly different approaches are the two sides of the same coin. What follows is just an early draft of my thoughts on this topic.