SIGNALS:EVOLUTION LEARNING & INFORMATION PAPER: Evolution, Learning, and Information

by Brian Skyrms

Brian Skyrms presents a fascinating exploration of how fundamental signals are to our world. He uses a variety of tools -- theories of signaling games, information, evolution, and learning -- to investigate how meaning and communication develop. He shows how signaling games themselves evolve, and introduces a new model of learning with invention. The juxtaposition of atomic signals leads to complex signals, as the natural product of gradual process. Signals operate in networks of senders and receivers at all levels of life. Information is transmitted, but it is also processed in various ways. That is how we think -- signals run around a very complicated signaling network. Signaling is a key ingredient in the evolution of teamwork, in the human but also in the animal world, even in micro-organisms. Communication and co-ordination of action are different aspects of the flow of information, and are both effected by signals.

Genres: Philosophy, Language, Science, Biology, Nonfiction, Linguistics, Evolution

207 pages, Paperback

First published April 1, 2010

Reviews

[Katja · Rating 4 out of 5]

Signaling games--the subject of this book--was a new topic to me. The book has less than 200 pages but the rate of how much you learn per page is impressively high (given that you knew very little in the beginning). The presentation is very accessible, so it is not that you need to consult Wikipedia or spend hours with pen and paper to make sense of it. The book starts with the simplest game possible which is gradually modified to illustrate different phenomena like deception, category formation, synonymy and ambiguity, emergence of logic and compositionality. Every chapter discusses how signaling evolves in different games, how signals can be learned, what are the equilibria in those games, etc. Skyrms style reminded me of simple Wikipedia: most sentences are shorter than ten words.

The reason for why I cannot give the book five stars is very poor editing, an embarrassment for Oxford UP. There are typos, unreferenced figures, wrong order in the index (check letter 'p'), very ugly, inconsistent formula formatting. For example, conditional probability, like p(x|s), is expressed as pr_s(x) in one place, pr(x given s) in another. Probability of x, p(x) is sometimes p_pr(x) and sometimes Probability(x). A times B is sometimes A*B and sometimes AB. This is annoying, clearly the fault of the editor. Finally, I was puzzled by the picture on the cover until I checked the title. Perhaps not the best choice for a book on signaling.

[David · Rating 4 out of 5]

Very interesting book on the spontaneous emergence of signalling systems.

[Steven · Rating 2 out of 5]

Muh culture replicators, Pythagorean goofiness.. blah blah blah. That's enough time on this.

[Zhiyi Li · Rating 4 out of 5]

I guess when you want to build "emergence" from fundementals, quite a lot can only be eluded to rather than gradually shown. In other words, on one hand you abstrat the basic elements, on the other hand you see the amazing emergent phenomenon. You can't really bridge the middle.

[aqeel · Rating 2 out of 5]

يتحدث عن تطور التعليم المُعزز، كثير من الفرضيات لم تكن منطقية بالنسبة لي.
Speaks about the evolution of reinforcement learning. Many theories used wasn't convincing.

[Nia Nymue · Rating 2 out of 5]

Sigh. He's got some really nice ideas here, really cool, but I dislike his style of writing. He seems to touch and go on some points without elaborating sufficiently - I'd think that he's going to talk in greater length and depth about something, but it stops suddenly and something new is brought up, making the points seem so underdeveloped. Disappointing. And he doesn't explain his diagrams - that bothers me a lot. There are also plenty of editing errors, as some other reviews have pointed out. If you are interested in signaling games, this book really is very interesting and recommended. But take note that it has many flaws. (It's hard to say if there are flaws in the arguments presented because they weren't presented well enough.)

[Alexander Telfar · Rating 2 out of 5]

read the first few chapters, then just skimmed the rest.
the writing isn't great but the topic seems interesting. I think I'll go elsewhere.
if you don't already know game and info theory then this book might be more interesting, or harder to read...