Do the right thing - studies in limited rationality by Stuart Russell and Eric Wefald, MIT Press, Cambridge, MA, £24.75, ISBN 0-262-18144-4

Chapter.6, on reasoning about actions, is not much better. Thanassas tries to distinguish between facts and events, and exemplifies the former by loud(noise, t\) (which is presumably meant to say that there is a loud noise at time t\, though the only way I can interpret it is as saying that the entity "noise" was loud at t\, which is rather different), and the latter by, inter alia, love(x, y, t\, t2). Of the latter he says that "since if x loves y in [t\ . . . t2] thenx loves y at any time point t between t\ and (2, this event conforms to the definition of a fact but does not conform to the definition of an event", which makes me wonder why he wants to classify it as an event in the first place. The confusions pile up thick and fast. Thanassas speaks of "the fact loud (noise, t\)" but he also refers to a fact's "changing its truth value". How can the fact that there was a loud noise at t\ change its truth value? This is what Quine (1960) calls an eternal sentence: its truth value does not depend upon the time at which it is uttered. Some of the problems arise, perhaps, from thinking that there is a meaningful dichotomy between facts and events. To my mind, the significant distinction is between states and events; and both of these can participate equally in facts, for example the fact that it is now dark (a fact about a state) and the fact that the sun set an hour ago (a fact about an event). At this point I did not see much point in reading further. I hesitate to condemn the substance of the book as worthless, although as I have indicated it contains many gross conceptual confusions. The problem is that the material is so badly presented that it is sometimes impossible to see what the author is trying to say. There are faults both in the way the material is organized and at the level of detail; there are distracting misprints and spelling errors on every page. This reflects badly on author and publisher alike. Indeed, I believe that the publisher should accept the major part of the blame on this score: the author is evidently not a native English speaker, and is inexperienced in publishing. Yet there is no sign that the book has been read through by a copy editor: the impression one has is that the author has submitted camera-ready copy and the publisher has printed it without further ado. Whether I am right or wrong in this surmise makes little difference: a book with this many errors should never have been published, and it is surely in the publisher's own interests to impose some more rigorous quality control than appears to have been exercised in this case. Speaking of errors, and returning now to the material itself rather than the presentation: who originated this nonsense about Australian penguins being able to fly (p.42)? How can a scientific researcher seriously work with a sentence such as "red birds are usually Australian penguins"? Of course, these are only light-hearted illustrative examples and not really about penguins. But one of these days, someone is going to read this sort of stuff and come away believing that Australian penguins really do fly, or that they are red. There is an intellectual irresponsibility here that should not be allowed to pass without comment. It betrays a cavalier approach to reality which indicates that its perpetrator is working in a fantasy world where real facts do not matter: it is all just an intellectual game.