Timing in the written production of German compounds

Presentation held at the 25th Annual Meeting of the German Society for Linguistics (DGfS) in Munich, February 26-28, 2003: "Sprache - Wissen - Wissenschaft". Workshop: Explaining Productivity, Chair: Peter Bosch & Anke Lüdeling

Guido Nottbusch & Ruediger Weingarten: Timing in the written production of German compounds

Abstract: The investigation of written word production gives many insights in processes that occur whilst a word is being written. Especially in morphologically complex words the relationship between compositional word production and whole retrieval can be clarified.

In previous studies (Will et al., 2002, Will et al., under review) we found that the time course of typing words correlates closely with their linguistic structures: The duration of inter-keystroke intervals (IKIs) was significantly longer when a syllable commenced with the following key and was longest when the key at the same time marked the beginning of a new morpheme (SM-type IKIs). The latter were influenced by word frequencies, indicating lexical dependencies. These effects were stable under various conditions (written, oral and pictorial stimulus) and in different languages (English, German). Therefore we conclude that the input into the motor system is composed of sub-word units instead of fully specified words. Accordingly this methods allows to investigate productive processes during word writing. We will here present the results of experiments investigating the influence of different types of morphemes on the time course of word production.

In Experiment 1 we compared in German words (a) prefix onsets (e.g. <vor+ent+halten> [to deprive] and stem onsets (e.g. <Meer+enge> [strait] and (b) suffix onsets (e.g. <Heiter-keit> [cheerfulness]) and again stem onsets [e.g. <Wasser-krug> [water jug]). The results yielded significant differences in delays only between stem and suffix onsets. This is discussed in a frame and filler concept, assuming that stems and prefixes open a new subword frame, whereas suffixes belong to the same subword frame as the preceding stem.

In Experiment 2 we examined the recent results concerning the differences between absolute and relative word frequencies in relation to productivity (Hay, 2000). Therefore we selected German compounds and prefixed words that have (a) a high surface frequency of the complex form but low relative frequency of the second part (e.g. <Gast-geber> [host]), possibly indicating that these words are retrieved as a whole, resulting in relatively short IKIs. Furthermore we contrasted the measured SM-IKIs with (b) words with a low absolute frequency but a high frequency of the stem morpheme beginning at the SM-boundary (e.g. <Blut-gruppe> [blood group]), probably being composed on-line, leading to relatively long IKIs. Significantly increased latencies at morpheme onsets were found only in low frequency items with relatively higher base frequency. This is interpreted as an effect of compositional word production.

In a further study we investigated the relationship between semantic compositionality, word and subword frequencies and the time course of word production. Productivity is then discussed in a model of hierarchichally ordered word frames and segmental fillers.

Hay, Jennifer B. (2000). Causes and Consequences of Word Structure. Diss., Northwestern University.

Will, Udo, Weingarten, Ruediger, Nottbusch, Guido, & Albes, Christian. (2002). Linguistische Rahmen und segmentale Informationen bei der Einzelwortschreibung. Evidenzen aus Zeitstrukturen und Fehlerverteilungen. In: Habel, Christopher; Pechmann, Thomas, Sprachproduktion. Opladen: Westdeutscher Verlag.

Will, Udo, Weingarten, Ruediger, Nottbusch, Guido, & Albes, Christian (under review). Linguistic units, hierarchies and dynamics of written language production.