Although the author has a very intimate and conversational style--indeed he seems a pleasant individual and one feels he is talking right to the reader face to face--at times it becomes a bit too much. For instance there was a tendency to repeat key words to the point that one felt brain washed! I counted the use of "dreams" some 14 times on a single page (p. 53)--and least you feel I'm a little picky, the next page continued the pattern, using "dreams" or "dreaming" another 14 times. While the following page only exhibited 10 repeats, 9 of them were in the first paragraph of 11 lines. This gets a little burdensome. I also felt that Professor Flanagan tended to overuse the technique of rephrasing his statements for emphasis a little too frequently. Usually with the first colorful simile or metaphor I've gotten the picture and enjoy the cleverness. By the second I've definitely gotten the idea, but by the third I've skipped on to the next paragraph!
I was impressed with the author's very broad background in the social sciences. He seemed particularly well read among the philosophers. He was also quite current on some of the newer research on dreaming and the physiology of sleep. For instance, he mentions the possibility that dolphins and some birds may sleep with half the brain--something we might all find useful at times. This would definitely have survival value by virtue of an awareness of the approach of predators. The author's suggestion that it might have arisen among dolphins because the breathing process may be fully under voluntary control and a heavily sleeping dolphin might stop breathing seemed unlikely to me. It would be more so that a fully sleeping dolphin might simply sink and drown!
I found the newer information on REM and NREM sleep interesting. I, like many who have studied the subject in the past, believed that all dreaming occurred during REM. I also didn't realize that there were different types of dreams during REM and NREM. Most particularly the physiological data on the brain site activity during the two phases was a surprise to me. The location of the REM site in the brainstem would have suggested to me that this was the older, more primitive form of sleep, since this area of the brain is considered the older portion of the nervous system. This is apparently not the case, as studies of various animals with different levels of nervous function indicate.
The author's footnotes were very informative. I don't always read end of chapter notes, but one of them caught my eye about mid-point in the book, and I actually went back and read those from the earlier chapters. The bibliography is quite good, with volumes ranging in date from 1910 to 1998. There were a number I added to my own wish list for the future. Journal articles included were from "Philosophical Review," "Journal of Neuroscience," "Science," "Psychological Review," "Brain," "Trends in Neuroscience," "Nature," "Behavioral Brain Research," and so on, mostly from the 1990s. Many of these might be a little more intimidating for the average reader--many may simply be unavailable unless the reader has access to a university library--but for anyone doing a school research paper they might make an additional source for study and follow-up.
In building his case, Flanagan opens with a startling proviso. He resurrects Stephen Gould's outmoded analogy applying architectural terms to biological processes. Gould's famous [and fatuous] use of the spandrel - a triangular form resting on the curve of an arch, derives from a 1976 article. Flanagan uses the analogy to declare dreams as "spandrels" but follows two contradictory themes in expressing it. In one, the spandrel is an necessary part of the arch - true if the arch supports anything like an aqueduct or roadway. In the other, the spandrel is not an essential part of an arch - true if you simply build an arch that has no other role. This issue wouldn't be terribly important except that it's the essence of Flanagan's argument and why he makes it.
Flanagan is at some pains to show there's no apparent evolutionary role for dreams. Sleep, of course, is another matter. Flanagan shows how many "rest" functions occur during sleep, with reduced impact on both brain and body allowing restoration. These are clearly "adaptive" traits to help the body survive. What role does dreaming play, then? Flanagan uses Gould's arguments and tactics to rule out dreaming as an adaptation because he can perceive no reason for dreams' occurrence. Flanagan adopts another Gould phrase, "exaptation," a trait that emerges in the past in one role which changes over time to assume another. No "exaptive" role can be discerned for dreams either, according to Flanagan. With "exaptations," you never know what they are until they've proven their worth as adaptations. By Flanagan's reasoning, everything is a Gouldian "spandrel" until you can properly assess its adaptive worth - some time in the next million years or so. Like Gould in his original essay, Flanagan provides no evidence for his claim since there is no means to discern any.
Flanagan's style is impressive in most respects. His descriptions are clear and his thesis forcefully presented. Prose skills, however, don't replace evidence. He provides a perplexing disclaimer on why only his own dreams are offered as data. He stresses that he sought dream evidence from family and friends, but that all denied him permission. With the wealth of published dream examples in the literature, this singular approach borders on the astonishing. Although examples of particular dreams have but little bearing on his thesis, it remains puzzling why he fails to use them to bolster, or challenge for refutation, his own case. A provocative book in many ways, it will be a challenge to scholars in human cognitive studies. Recommended chiefly for the professional, it yet provides an entertaining, if not informative read.