The influence of right or left handedness on the ability to simulate handwritten signatures and some elements of signatures: a study of Arabic writers.

It has been widely assumed since the 1970s that right-handed writers, on average, do not write or simulate handwriting any better than left-handed writers. This study has dug deeper into that question, to find narrow language environments which left-handed and right-handed writers simulate with different degrees of success or making different types of errors. These might provide useful markers of handedness in writers or forgers. The sample of 823 native Arabic-writing adults, 763 right handed and 60 left handed, first attempted to simulate two Arabic signatures. As expected, the accuracy of the simulations of the two groups was not significantly different. Simulation accuracy of a variety of narrower elements was then measured to search for small environments in which the two groups performed differently. Two elements - slant and alignment - were significantly more poorly simulated by left-handed subjects than right-handed subjects, although the differences were not great enough to be useful in suggesting the handedness of a forger. These differences in simulation accuracy were not totally explored or explained. Instead, the differences in types of errors to slant and to alignment were investigated. Left-handed simulators, but not right-handed simulators, showed a strong tendency that had been predicted for both groups: to slant strokes and letters toward the side of the writing hand, when the target strokes and letters slanted in the opposite direction. The difference between the handedness groups was great enough that a strong preference for leftward slant in an Arabic writer or forger might suggest left handedness, but not, of course, rule out right handedness. Right-handed simulators, but not left-handed simulators, showed a strong tendency that had also been predicted for both groups: for the line of writing to be higher on the side where the writing hand was located. Again, the difference between handedness groups was great enough that a strong tendency for the writing line to fall to the left in Arabic writing and simulation might suggest right handedness, but would certainly not rule out left handedness. Thus, suggestive, but not absolute, markers of handedness have been found. In an effort to explain these almost mirror-image patterns, it was noticed that the operation of two more tendencies, in addition to the operation in both groups of the two predicted tendencies, might explain the patterns of the data. If we assume that both handedness groups tended to slant toward the writing hand and tended to make the writing line higher on the side of the writing hand, the two other tendencies operating in both groups would have been a tendency to slant strokes and letters toward the end of the writing line, and a tendency for writing to fall toward the end of the writing line. The operation of these four tendencies would account quite well for the right-handed and left-handed patterns in the two groups.