Though it's difficult to agree on the exact date of their union, logic and artificial intelligence (AI) were married by the late 1950s, and, at least during their honeymoon, were happily married. But what connubial permutation do logic and AI find themselves in now? Are they still (happily) married? Are they divorced? Or are they only separated, both still keeping alive the promise of a future in which the old magic is rekindled? This paper is an attempt to answer these questions via a review of the following six books:
One side-effect of the investigation undergirding our answer is a taxonomy for parsing logicist AI and, possibly, the start of a taxonomy for parsing connectionist AI as well. We also intend the following to be useable as an introduction to logicist AI (for readers with a modicum of previous exposure to logic and computer science), as an antidote to caricatures of logicist AI,It is sometimes said that the logicist approach to AI is the ``sentence view of the mind.'' Given the mathematical maturity that logicist AI has reached--maturity we seek to convey herein--this is like viewing neural nets as merely strings of Christmas tree lights and calling connectionist AI the ``little light bulb view of the mind.'' and, to some degree, as a rallying cry for the logicist approach to understanding and replicating (or at least simulating) the human mind.
We should note at the outset that the marriage in question may not be monogamous. The connectionist would claim that AI has been polygamous from the outset, that AI was long ago married to non-logical (= sub-symbolic) architectures and approaches (yielding CAI, for connectionist AI; pronounced so as to rhyme with `fly'), that a separation took place (courtesy, so the tale goes, of Minsky and Papert), and that the passion returned starting in the early 1980s. The CAInik would also claim, no doubt, that this marriage will eventuate in intelligent offspring of a type beyond the reach of LAI (pronounced--gleefully, by the connectionist--as `lie'), the logic and AI couple. The second of these claims, for the record, is one we reject, but we make no effort to defend our rejection herein.We have elsewhere (Bringsjord 1991) specified and defended the view that if AI is to succeed in building persons, it must indeed be polygamous. In a word, our view is that CAI on its own will produce, at most, a mere animal incapable of the sorts of behaviors (e.g., belief fixation on the basis of technical philosophical dialectic) that at least appear to be symbolic in nature. LAI, on the other hand, will, if pursued alone, at most produce what John Pollock (1995) has called `artilects', beings capable of ratiocination but incapable of, say, playing baseball. That such artilects could in principle be persons, even without any sensory interchange with the environment, is a claim defended in Bringsjord & Zenzen 1991.