The syllable-length effect in number processing is task-dependent

Abstract

Two experiments were run in order to investigate the relationship between syllable length of number names and eye-fixation durations during silent reading of one-and two-digit numbers. In Experiment 1, subjects had to read a series of three numbers and recall them orally; in Experiment 2, subjects had to indicate manually whether the value of the middle number was between the values of the outer numbers. The effect of syllable length was controlled for possible confounding effects of number frequency and number magnitude. Findings indicated that fixation duration depended on syllable length of number names in the first task, but not in the second task. The results call into question the claim that phonological encoding is imperative in visual processing; phonological encoding was used only when the numbers had to be recalled, and not when they were coded for computational purposes.Pynte (1974) showed that the relationship between syllable length of number names and response latency previously found for number naming (Eriksen, Pollack, & Montague, 1970;Klapp, 1971Klapp, , 1974) also existed in silent number reading. Two series of three two-digit numbers were each presented for 2 sec. Subjects had to read the numbers silently and to report them aloud as soon as the stimulus display disappeared. Eye movements were tracked while the subjects were reading. Pynte (1974) found that gaze duration on the center number increased as the syllable length of the number name increased, although pronunciation was not required during processing. This led Pynte to the conclusion "that a mechanism of readiness for pronunciation co-occurs with visual processing" (p. 110).Pynte's (1974) finding is regularly cited, along with other studies, as evidence that phonological encoding is an essential element in visual processing (see, e.g., Hung & Tzeng, 1981; Rayner & Pollatsek, 1989, p. 214;Rossi, 1985). However, before Pynte's (1974) results can be used as unequivocal support for phonological mediation in reading, alternative interpretations must be ruled out. Pynte used only two times two critical stimuli: "30" (trente, one syllable) versus "90" (quatre-vingt-dix, four syllables), and "28" (yingt-huit, two syllables) versus "82" tquatre-vingt-deux, four syllables). This leaves open the possibility that the findings were due to factors other than syllable length of the stimulus name. First, Pynte did not consider the frequency of occurrence of his stimuli, so that his results may have been an artifact of the large num- bers' having lower frequencies than the low numbers. Second, it is not unreasonable to assume that large numbers are more difficult to process than small numbers, so that Pynte's syllable effect might also have been due to differences in number magnitude. Finally, it can be asked whether Pynte really did design a task •'without pronunciation. " Though naming was not required during reading, the subsequent oral report of numbers may have entailed subvocalization as part of response preparation.To address the three crit...

A neural model of how the brain represents and compares multi-digit numbers: spatial and categorical processes

“…10) that was similar to gaze duration times for single digit reading (Gielen, Bryzbaert, & Dhont, 1991). Some psychophysical studies have reported logarithmic behavior with respect to number magnitude, such as subitizing experiments (Mandler & Shebo, 1982) and two-digit number reading experiments (Bryzbaert, 1995).…”

“…In the experiments by Gielen, Bryzbaert, and Dhont (1991), the gaze duration is used to estimate the time interval needed by the subjects to process a numerical stimulus under no time stress. This interval ranged from 200 to 600 ms for one-digit and two-digit numbers.…”

“…Fig. 10(diamonds, dashed line) displays the first-gaze duration (FGD) times in Arabic number reading experiments (Gielen, Bryzbaert, & Dhont, 1991). Here, the Number Size effect yields an approximately linear increase in reading time with number magnitude.…”

Incrementality in naming and reading complex numerals: evidence from eyetracking

“…The regressions indicated that the number of syllables and the numeral frequency, but not the number magnitude, determined the gaze durations, even when the contribution of the other factors was partialled out. However, the results by Gielen et al (1991) were contested in a follow-up study by Brysbaert (1995). When controlling for several methodological factors, such as lack of power and inclusion of the numeral 0, Brysbaert (1995) showed that gaze durations for numerals ranging from 1 to 99 were a function of the logarithm of magnitude.…”

“…Pynte's stimuli differed not only in number of syllables but also in number magnitude and numeral frequency. In order to determine the contribution of all three factors, Gielen, Brysbaert, and Dhondt (1991) conducted multiple regression analyses on the gaze durations in the silent naming of all one-and two-digit arabic numerals ranging from 0 to 99. The regressions indicated that the number of syllables and the numeral frequency, but not the number magnitude, determined the gaze durations, even when the contribution of the other factors was partialled out.…”