Marian Annett's recent book on handedness and cerebral dominance was published a dozen years after Efron's (1990) declaration that the era of hemispheric specialization had ended. Why did that era not end for Annett? One answer is that Annett's work differed from that of the majority of laterality researchers. Even during the heyday of laterality studies, Annett remained aloof from the mainstream endeavor of characterizing left- and right-hemisphere functions. Instead, Annett pursued a line of laterality research that dealt primarily with handedness in the general population. She accumulated large quantities of data about human handedness and used those data to support her right-shift (RS) theory of the genetics of handedness. With the publication of her new book, the interested reader has convenient access to the fruits of a remarkably focused and durable research career.

Annett describes her book as a detective story. The story begins with the author's skepticism concerning the “many layers of opinion and surmise” that constituted the older literature on handedness and brain asymmetry. Starting over, she proceeded to build a new theory that was grounded in quantitative relationships found in large samples. Relying more on biology than medicine to guide her thinking, she constructed a large corpus of work that she presents coherently in her book.

The first chapter is a fine introduction to the “puzzle” of handedness and its relation to language lateralization. This chapter by itself would serve well as an assigned reading for a graduate class on laterality. In Chapter 2, after arguing that hand preference is a continuous variable, Annett explains how she used the botanists' technique of association analysis to develop and validate the 12-item Annett Hand Preference Questionnaire (AHPQ). The AHPQ, along with a peg-moving task, accounts for much of Annett's handedness data.

Chapters 3–5 constitute the theoretical core of the book, for it is here that Annett describes the fundamental tenets and implications of the RS theory. In these chapters, the author guides the reader step by step through the observations that underlie the RS theory. Once Annett understood that handedness is a continuous rather than a discrete variable, it became evident to her that the proportions of right-, mixed- and left-handedness form a binomial distribution with respective percentages of 66, 30 and 4. The data on hand and paw preference in nonhuman animals also conform to a binomial distribution, but the distribution is symmetrical. Then came Annett's eureka! experience, her recognition that human handedness is animal pawedness with the addition of a right-shift (RS) factor. Or, to state the RS theory more carefully: Human handedness is the result of chance plus a species-specific bias toward dextrality (the RS factor) and cultural enhancement of dextrality. Annett uses published compilations of aphasia cases to estimate the population frequency of the right-shift gene (RS+) and to specify the relationship between speech lateralization and handedness. Then she demonstrates that a single gene can account for patterns of handedness in families and in pairs of monozygotic and dizygotic twins.

In the next part of the book (Chapters 6–9), Annett addresses various topics: stability of handedness, implications of an additive version of the RS model, asymmetries other than handedness and speech lateralization, and associations among the different asymmetries. A principal conclusion is that, irrespective of the asymmetry in question, about two-thirds of individuals show a lateral bias (e.g., toward left-sided language) whereas the remaining one-third show a random distribution of asymmetries around an unbiased mean.

As if to accommodate the reader who is beginning to grow weary of handedness studies, the next section of the book (Chapters 10–14) entails a plot twist that expands the scope of the RS theory and makes it relevant to phenomena outside the realm of laterality. In this section, Annett articulates her thesis that the three genotypes (RS++, RS+− and RS−−) exemplify a balanced polymorphism with heterozygote advantage. If the RS+ allele were entirely advantageous to its bearer, she reasons, people with the RS++ genotype should have become ubiquitous. The high prevalence of heterozygosity (as inferred from handedness distributions) suggests that both benefits and costs come with the RS+ allele. Annett argues that the primary benefit is excellence in perceiving and producing speech sounds, whereas the primary cost is relatively poor skill with the left hand. Both characteristics are derived from the effect of the RS+ allele on the left cerebral hemisphere. The reader is then told about a far-ranging assortment of studies—studies of cognitive abilities, educational attainments and prowess in tennis and cricket, as well as studies of people with learning disabilities, schizophrenia and autism—that the author offers as support for the benefits of heterozygosity.

In the final section (Chapters 15–17), Annett discusses a few independent tests of the RS theory and briefly reviews alternative theories of handedness. She points out ways in which the RS theory differs from the genetic theory of McManus (1985), with which it is sometimes confused. She succinctly states the principal tenets of RS theory in the last chapter before discussing some implications of the theory for psychological processes and offering some speculations about the evolution of speech and laterality.

A detective, scientific or otherwise, has a choice of investigative styles. She may adopt an inductive style that requires the painstaking accumulation of clues before a hypothesis is selected. Alternatively, she may choose a deductive style in which a hypothesis is formulated early in the investigation and then used to guide the search for more evidence. Annett has embraced the latter style even though this choice may cause the evidence to be seen through a selective filter. Evidence that is incompatible with the hypothesis may be overlooked and ambiguous evidence may be interpreted in a way that favors the hypothesis.

Indeed, there are numerous instances in which Annett makes questionable post-hoc assumptions and adjustments to reconcile evidence with her theory. One example pertains to the handedness of adults who have become dysphasic following stroke. An elevated frequency of left-handedness typically is taken as evidence that language is more often represented bilaterally in left-handers than in right-handers, thus making left-handers more susceptible to at least a transitory dysphasia following damage to either hemisphere. Annett, however, after pinpointing data that are inconsistent with RS theory, concludes that the investigators must have used different criteria for classifying the handedness of dysphasic and non-dysphasic patients. Another assumption about handedness criteria is invoked to bring the results of sodium Amytal testing into agreement with the RS theory. Additional post-hoc adjustments include reductions in the expressivity of the right-shift gene in twins and the introduction of “pull to concordance” corrections to accommodate data on the joint distributions of writing dominance with throwing and kicking dominance.

By its nature, the RS theory offers an abundance of interpretative latitude. Because the suspected RS gene has not been identified, the three genotypes cannot be distinguished on the basis of genetic testing. The probabilities of the different genotypes within a sample must be inferred from the distribution of relevant phenotypic characteristics such as handedness. However, because handedness is determined primarily by chance, there is only a weak relationship between handedness and the putative genotype. The problem is complicated by the fact that handedness is defined as hand preference in some studies and as asymmetry of manual skill in other studies. The two operational definitions are correlated but not identical. For all those reasons, the predictions of RS theory are difficult to confirm or disconfirm unequivocally.

The student of handedness is advised to complement Annett's book with other readings that reflect different perspectives. This comment is not intended to impugn the merits of RS theory, but only to point out that the author is an unabashed advocate of the theory. Annett is a clever detective, but she never loses sight of what she wants to find. As she confesses at one point, readers should not expect her to be impartial. Whereas her confession may have been whimsical, it raises some serious questions about the role of theory in behavioral science. What is the purpose of a theory? Must a theory be correct if it is to have scientific value? What is a reasonable expectation for support of a theory? Should one demand a perfect fit? An approximate fit? A mixture of good and poor fits if the poor fits can be patched up by post hoc assumptions? Can the creator of a theory ever be an impartial arbiter of the evidence, or should behavioral science be divided—following the example of physics—into theoretical and experimental branches?

The book is well written and, given the influence and scope of RS theory, it is essential reading for any neuropsychologist with a strong interest in laterality. However, debate about this particular detective story will not end here. As much as Annett would like her right-shift theory to be the starting point for the ultimate explanation of human handedness, it is too soon to know. The answers may have to await discoveries at the genetic level of analysis. In the meanwhile, the theory's correctness may be less important than the heuristic value of some of its component concepts, such as the prominent role of chance and the proposition that handedness is derived from speech lateralization (rather than vice versa).

So long as most human beings write and throw with the right hand and use the left side of the brain to control speech, the phenomenon of hemispheric specialization will continue to attract researchers. As new evidence becomes available, Annett, no doubt, will interpret that evidence in terms of her theory and others, no doubt, will disagree. Maybe the era of hemispheric specialization is only beginning.