Are long compound words identified serially via their constituents? Evidence from an eyemovement-contingent display change study

Abstract

In our prior research, we examined the identification of two-noun Finnish compound words by embedding them in sentence contexts and recording readers' eye movements on these compounds while they read the sentences for comprehension. (In Finnish, compound words are never broken by interword spaces.) The results suggest that long (12-to 18-letter) compound words are processed primarily in such a way that the constituents are processed serially (from left to right). Hyönä and Pollatsek (1998) manipulated the frequency of the first constituent (i.e., its frequency as a separate word) while matching for second-constituent and whole-word frequency (e.g., liikenne in liikennevalo ϭ traffic light). They observed that first-constituent frequency had both an early effect that surfaced in the duration of the initial fixation on the word and a later, larger effect on the gaze duration on a word (the total first pass time on the word), which was due largely to differences in the number of times that the word was fixated. Both of these effects were subsequently replicated by Bertram and Hyönä (2003) for long compounds. Pollatsek, Hyönä, and Bertram (2000) also observed that the frequency of the second constituent had a large effect on the gaze duration on the compound word. However, second-constituent frequency did not affect the duration of first fixation on the compound but affected processing starting only on the second fixation. These data are consistent with the view that long compounds are identified serially via their constituents by first accessing the initial constituent, followed by the access of the second constituent (for other eye movement studies of compound word processing in reading, see Andrews, Miller, & Rayner, 2004;Inhoff, 1989a;Inhoff, Briihl, & Schwartz, 1996;Inhoff, Radach, & Heller, 2000;Juhasz, Starr, Inhoff, & Placke, 2003). However, since Pollatsek and Hyönä (in press) demonstrated that both the gaze durations and the first-constituent frequency effects were the same for semantically opaque compounds as for semantically transparent compounds, it appears that the initial stages of the composition of the two constituents does not involve integrating their meanings.There are data, however, indicating that accessing a compound word is more than a serial identification of its components. In particular, when the frequency of the compound word was varied (while equating the frequency of the constituents), there were also large effects on the gaze duration on the compound word , although the effect of word frequency was not significant until the second fixation. This word frequency effect indicates that more holistic processing of the compound words is occurring as well. Moreover, Bertram and Hyönä (2003) found that a compositional mode of processing may be important only for relatively long compound words. For shorter compound words (7-9 characters), they obtained an early effect of word frequency but 523Copyright 2004 Psychonomic Society, Inc.This research was supported by an Academy of Finland grant to t...

Parafoveal processing in reading

“…To what extent do word length and the spaces between words and within spaced compound words affect the extent to which they are processed parafoveally? Hyönä, Bertram, and Pollatsek (2004) conducted a boundary experiment in which the boundary was located between two morphemes constituting an unspaced Finnish compound word. They found large preview benefits on the second morpheme, yet the availability of preview of the second morpheme did not affect fixation times on the first morpheme.…”

Reading polymorphemic Dutch compounds: Toward a multiple route model of lexical processing.

“…For instance, Hyönä et al (2004) found that for long compounds there is early activation of the left constituent (dish) and later activation of the right constituent (washer). However, two important questions about the time-course of morphological processing are as yet unresolved.…”

“…Whatever the disadvantages of our methodology may be, the pattern of results that we have obtained and reported either in the body of the paper or in Appendix 1, dovetails perfectly with many of the results obtained in the literature for sentential reading, such as visuo-oculomotor effects (cf., e.g., O'Regan et al, 1994;Rayner, 1998;Vitu, McConkie, Kerr & O'Regan, 2001), effects of compound length and frequency, as well as of constituent frequencies (cf., e.g., Andrews et al, 2004;Duñabeitia, Perea & Carreiras, 2007;Hyönä & Pollatsek, 1998;Hyönä et al, 2004;Juhasz et al, 2003;Taft & Forster, 1976), and effects of orthographic n-grams (reported in Appendix 1, cf., e.g., Lima & Inhoff, 1985). Furthermore, in a recent sentential reading study (Kuperman, Bertram & Baayen, 2008), in which Finnish compounds were embedded in context, a highly similar pattern of results was observed, including early effects of compound frequency, left constituent frequency and family size, later and weaker effects of right constituent frequency and family size, interactions between morphological predictors, as well as longitudinal experimental effects.…”

Parafoveal processing across different lexical constituents in Chinese reading

“…Although traditionally parafoveal-on-foveal effects have been examined between adjacent words in a sentence, more recently there have been a number of experiments employing the boundary paradigm within words to investigate whether the constituents of morphologically complex words (e.g., compound words like basketball) are processed serially or in parallel (e.g., Drieghe, Pollatsek, Juhasz, & Rayner, 2010;Hyönä, Bertram, & Pollatsek, 2004;Juhasz, Pollatsek, Hyönä, Drieghe, & Rayner, 2009). That is to say, these experiments have investigated serial or parallel lexical processing of linguistic constituents within a word rather than between words.…”

“…That is to say, these experiments have investigated serial or parallel lexical processing of linguistic constituents within a word rather than between words. Hyönä et al (2004) used the boundary paradigm to investigate processing of Finnish sentences containing 12-to 18-letter compound words. The boundary was located between the first and second constituent of the compound, and the preview of the second constituent was manipulated (being a visually dissimilar letter string, or an identity string).…”