The interaction of central and peripheral processes in typing and handwriting: A direct comparison.

Abstract

This study aimed to investigate the interaction between linguistic and peripheral-motor processes in written production. Past research has focused on this topic by analyzing how handwriting and, more recently, typing execution were influenced by lexical and sublexical variables. We took a step further in this study by directly comparing handwriting and typing, examining if different motor executions allow for different flows of linguistic processing. Participants typed and handwrote a set of Italian stimuli in which we manipulated lexicality (words vs. pseudowords), orthographic complexity (stimuli with vs. without multiletter graphemes), and length (short vs. long stimuli). We measured and analyzed latency (response times [RTs]), the difference between RTs and the acoustic duration (AD) of the stimuli (RT-AD), mean length of interletter intervals (ILIs), and whole response duration (WRD). We further explored the effects of the position of the orthographic complexity on RTs, RT-AD, ILIs, and WRD. Results suggested a cascaded, continuous processing flow for handwriting and a mixed mechanism involving both serial and parallel modes of processing for typing. The differences in linguistic processing during handwriting and typing suggest different mechanisms in segmenting, maintaining, and retrieving the orthographic representation during motor execution. Public Significance StatementThis study suggests that handwriting and typing allow for different processing of linguistic information during their execution. Linguistic information, such as length, orthographic complexity, and lexicality, affect handwriting movements. During typing, the length of the stimulus has a strong influence, while most of the lexicality and orthographic complexity processing is moved up at the beginning of the stimulus execution (before typing starts and during first key presses). The differences in linguistic processing during handwriting and typing production suggest differences in preparing and maintaining in memory the to-be-written sequence.

Spelling monosyllabic English words

“…Note that it might not be possible to distinguish between serial and parallel models because certainly there will be a length effect in the duration of the response, but we will not know if it is due to computation of the orthographic code or the implementation of the motor code. We note that with Italian words that varied from five letters and two syllables to nine letters and four syllables, Cerni and Job (2022) found that response execution in both handwriting and typing began especially early for the longer items.…”

“…Perhaps more interesting is the evidence that the written response appears to cascade over time such that the processing of central lexical access and graphemic construction processes may not be fully complete when the execution of the peripheral motor response (i.e., of typing or writing) commences. This highlights the utility of collecting both measures of the response initiation as well as the duration of the spelling response over time (e.g., Cerni & Job, 2022;Roux et al, 2013) We note that our first key RT measures the time between the offset of the acoustic signal and the first keypress. This procedure differs from other spelling studies that measure RT from the onset of the acoustic signal (e.g., Bonin et al, 2015).…”

Carport and carpet: Effects of compound and pseudocompound word structures on typing.

“…Although there is evidence that linguistic factors can affect typing performance (see previous section), the question of serial versus cascaded processing has not been fully resolved. Indeed, a recent article examining handwriting and typing concluded: “For handwriting, the pattern of results was compatible with a cascaded flow of linguistic processing during written execution, while for typing, a more intertwined pattern of cascade and serial processing emerged” (Cerni & Job, 2022, p. 576). One potential explanation for this variability in results is that the morphological structure of a word might influence the impact of lexical information, especially of word frequency.…”

“…Latencies for the initial letter of the word are faster for high-frequency words than for lower-frequency words (Baus et al, 2013;Feldman et al, 2019;Inhoff, 1991;Pinet et al, 2016;Torrance et al, 2018). In contrast, although there has been some evidence suggesting that some latencies at noninitial positions are affected by word frequency (Gentner, 1983;Gentner et al, 1988), the majority of studies have found that word frequency did not influence typing latencies for noninitial letters (e.g., Baus et al, 2013;Bertram et al, 2015;Cerni & Job, 2022;Pinet et al, 2016). Thus, it appears that higher word frequency has a facilitatory effect at the word level (e.g., initiating the first letter) but does not affect the letter-output level (e.g., the latencies for the noninitial letter keystrokes) or is more difficult to detect at noninitial positions.…”

“…An important theoretical question stemming from differences among the various frameworks of typing production is which types of lexical and sublexical information (if any) are involved in the planning and execution of keystrokes (e.g., Cerni & Job, 2022;Feldman et al, 2019;Massaro & Lucas, 1984;Pinet et al, 2016). This is a particularly relevant question because, as noted in the previous section, the various frameworks make a range of predictions concerning the involvement of this information, especially in terms of sublexical information with some (e.g., Logan, 2018) positing that this information is not relevant for motor output and others (e.g., Gagné & Spalding, 2014c, 2016bLibben et al, 2012;Weingarten, 2005) positing that sublexical units are important planning units for output.…”