The balance of storage and computation in morphological processing: the role of word formation type, affixal homonymy, and productivity

This article is concerned with the way in which the balance of storage-storing and processing words through full-form representations-and computation-storing and processing words through morpheme-based representations-in lexical processing in the visual modality is affected by the following 3 factors: word formation type (roughly, inflection vs. derivation), productivity, and affixal homonymy. Experimental results for 5 different Dutch suffixes, combined with previous results obtained for 4 comparable Finnish suffixes (R. Bertram, M. Laine, & K. Karvinen, 1999) and 2 Dutch suffixes (R. H. Baayen, T. Dijkstra, & R. Schreuder, 1997), show that none of these factors in isolation is a reliable cross-linguistic predictor of the balance of storage and computation. The authors offer a general framework that outlines how morphological processing is influenced by the interaction of word formation type, productivity, and affixal homonymy.     

An amorphous model for morphological processing in visual comprehension based on naive discriminative learning

A 2-layer symbolic network model based on the equilibrium equations of the Rescorla-Wagner model (Danks, 2003) is proposed. The study first presents 2 experiments in Serbian, which reveal for sentential reading the inflectional paradigmatic effects previously observed by Milin, Filipovi? ?ur?evi?, and Moscoso del Prado Martín (2009) for unprimed lexical decision. The empirical results are successfully modeled without having to assume separate representations for inflections or data structures such as inflectional paradigms. In the next step, the same naive discriminative learning approach is pitted against a wide range of effects documented in the morphological processing literature. Frequency effects for complex words as well as for phrases (Arnon & Snider, 2010) emerge in the model without the presence of whole-word or whole-phrase representations. Family size effects (Moscoso del Prado Martín, Bertram, H?iki?, Schreuder, & Baayen, 2004; Schreuder & Baayen, 1997) emerge in the simulations across simple words, derived words, and compounds, without derived words or compounds being represented as such. It is shown that for pseudo-derived words no special morpho-orthographic segmentation mechanism, as posited by Rastle, Davis, and New (2004), is required. The model also replicates the finding of Plag and Baayen (2009) that, on average, words with more productive affixes elicit longer response latencies; at the same time, it predicts that productive affixes afford faster response latencies for new words. English phrasal paradigmatic effects modulating isolated word reading are reported and modeled, showing that the paradigmatic effects characterizing Serbian case inflection have crosslinguistic scope.     

Affixal homonymy triggers full-form storage, even with inflected words, even in a morphologically rich language

This paper investigates whether affixal homonymy, the phenomenon that one affix form serves two or more semantic/syntactic functions, affects lexical processing of inflected words in a similar way for a morphologically rich language such as Finnish as for morphologically restricted languages such as Dutch and English. For the latter two languages, there is evidence that affixal homonymy triggers full-form storage for inflected words (Bertram, R., Schreuder, R., and Baayen, R. H. (in press). The balance of storage and computation in morphological processing: the role of word formation type, affixal homonymy, and productivity. Journal of Experimental Psychology: Learning, Memory, and Cognition; Sereno and Jongman (1997). Processing of English inflectional morphology. Memory and Cognition, 25, 425-437). Two visual lexical decision experiments show the same pattern for Finnish. Apparently, the substantially richer morphology in Finnish does not prevent full-form storage for inflected words when the affix is homonymic.     

The subjects as a simple random effect fallacy: subject variability and morphological family effects in the mental lexicon

This is a methodological study addressing the appropriateness of standard by-subject and by-item averaging procedures for the analysis of repeated-measures designs. By means of a reanalysis of published data (Schreuder & Baayen, 1997), using random regression models, we present a proof of existence of systematic variability between participants that is ignored in the standard psycholinguistic analytical procedures. By applying linear mixed effects modeling (Pinheiro & Bates, 2000), we call attention to the potential lack of power of the by-subject and by-item analyses, which in this case study fail to reveal the coexistence of a facilitatory family size effect and an inhibitory family frequency effect in visual and auditory lexical processing.     

The precise time course of lexical activation: MEG measurements of the effects of frequency, probability, and density in lexical decision

Visually presented letter strings consistently yield three MEG response components: the M170, associated with letter-string processing (Tarkiainen, Helenius, Hansen, Cornelissen, & Salmelin, 1999); the M250, affected by phonotactic probability, (Pylkk?nen, Stringfellow, & Marantz, 2002); and the M350, responsive to lexical frequency (Embick, Hackl, Schaeffer, Kelepir, & Marantz, 2001). Pylkk?nen et al. found evidence that the M350 reflects lexical activation prior to competition among phonologically similar words. We investigate the effects of lexical and sublexical frequency and neighborhood density on the M250 and M350 through orthogonal manipulation of phonotactic probability, density, and frequency. The results confirm that probability but not density affects the latency of the M250 and M350; however, an interaction between probability and density on M350 latencies suggests an earlier influence of neighborhoods than previously reported.     

汉语音节累积词频对同音字听觉词汇表征的激活作用

汉语同音字音节的累积词频是同音单字词的具体词频之和, 即音节作为词被听到的频数, 也称为音节词频。它对词汇通达的作用还没获得一致看法。本文探讨听觉词汇通达中同音字表征的激活如何受累积词频的影响, 构成累积词频的具体词频和同音字数又会在其中起何作用。研究采用跨通道的音-字同音判断范式和配套的协方差分析分离出听觉的词汇表征激活作用。实验1固定同音字数, 同时变化累积词频和具体词频(其中具体高、低频同时也是同音字族内的高、低频), 结果发现具体词频的提高对表征激活有积极作用; 累积词频的提高仅在错误率上产生微弱的抑制作用, 而这个微弱作用可以归结为同音字族内高频字表征的激活对低频字表征的抑制。实验2固定具体词频, 同时变化累积词频和同音字数, 结果发现同音字数增加对表征激活产生抑制, 没有发现累积词频的作用。两个实验的结果说明, 音节累积词频对同音字的听觉词汇表征激活不起确定作用, 能够起作用的是同音字族内的具体词频和同音字数; 但同音字数变化的抑制作用仅在具体词频不变时才发生, 因此起决定作用的是同音字族内高频字表征激活和低频字表征抑制的机制。     

Is Milkman a superhero like Batman? Constituent morphological priming in compound words

In the present study, we examined morphological decomposition of Basque compound words in a series of masked priming lexical decision experiments. In Experiment 1, Basque compound words could be briefly preceded by other compounds that shared either the first or second constituent, or by unrelated noncompound words. Results showed a significant priming effect for words that shared a constituent, independently of its position. In Experiment 2, compound words were preceded by other compound words that shared one of their constituents, but in a different lexeme position (e.g., the first constituent of the compound that acted as a prime was the second constituent of the compound that acted as a target). Results again showed a constituent priming effect (i.e., location in the string is not necessary for priming to occur). In Experiment 3, we demonstrated that these priming effects were not due to mere form overlap: pairs of noncompound words that shared either the beginning or the ending chunk did not produce a priming effect. Taken together, the present results converge with previous data on orthographic/morphological priming and provide evidence favouring early morphological decomposition.     

Naming the color of a word: is it responses or task sets that compete?

Subjects named the colors in which high- and low-frequency words and pronounceable nonwords, otherwise matched, were displayed. Color naming was slower for all three item types than for visually equivalent strings of nonalphanumeric symbols but was no slower for words than for nonwords, nor for high-frequency words than for low-frequency words. Unpronounceable letter strings had intermediate color-naming latencies. However, frequency and lexical status had large effects on latency for reading the same words and pseudowords aloud. Interference is thus predicted not by the strength of association between a letter string and its pronunciation but by the presence of word-like constituents. We argue that the interference from an unprimed noncolor word is due to, and isolates, one of two components of the classic Stroop effect: competition from the whole task set of reading. The other component, response competition, occurs only when lexical access is sufficiently primed.     

Morphological family size in a morphologically rich language: the case of Finnish compared with Dutch and Hebrew

Finnish has a very productive morphology in which a stem can give rise to several thousand words. This study presents a visual lexical decision experiment addressing the processing consequences of the huge productivity of Finnish morphology. The authors observed that in Finnish words with larger morphological families elicited shorter response latencies. However, in contrast to Dutch and Hebrew, it is not the complete morphological family of a complex Finnish word that codetermines response latencies but only the subset of words directly derived from the complex word itself. Comparisons with parallel experiments using translation equivalents in Dutch and Hebrew showed substantial cross-language predictivity of family size between Finnish and Dutch but not between Finnish and Hebrew, reflecting the different ways in which the Hebrew and Finnish morphological systems contribute to the semantic organization of concepts in the mental lexicon.     

Probability and surprisal in auditory comprehension of morphologically complex words

Two auditory lexical decision experiments document for morphologically complex words two points at which the probability of a target word given the evidence shifts dramatically. The first point is reached when morphologically unrelated competitors are no longer compatible with the evidence. Adapting terminology from Marslen-Wilson (1984), we refer to this as the word's initial uniqueness point (UP1). The second point is the complex uniqueness point (CUP) introduced by Balling and Baayen (2008), at which morphologically related competitors become incompatible with the input. Later initial as well as complex uniqueness points predict longer response latencies. We argue that the effects of these uniqueness points arise due to the large surprisal (Levy, 2008) carried by the phonemes at these uniqueness points, and provide independent evidence that how cumulative surprisal builds up in the course of the word co-determines response latencies. The presence of effects of surprisal, both at the initial uniqueness point of complex words, and cumulatively throughout the word, challenges the Shortlist B model of Norris and McQueen (2008), and suggests that a Bayesian approach to auditory comprehension requires complementation from information theory in order to do justice to the cognitive cost of updating probability distributions over lexical candidates.     

Does word frequency affect lexical selection in speech production?

We evaluated whether lexical selection in speech production is affected by word frequency by means of two experiments. In Experiment 1 participants named pictures using utterances with the structure “pronoun + verb + adjective”. In Experiment 2 participants had to perform a gender decision task on the same pictures. Access to the noun's grammatical gender is needed in both tasks, and therefore lexical selection (lemma retrieval) is required. However, retrieval of the phonological properties (lexeme retrieval) of the referent noun is not needed to perform the tasks. In both experiments we observed faster latencies for high-frequency pictures than for low-frequency pictures. This frequency effect was stable over four repetitions of the stimuli. Our results suggest that lexical selection (lemma retrieval) is sensitive to word frequency. This interpretation runs against the hypothesis that a word's frequency exerts its effects only at the level at which the phonological properties of words are retrieved.     

Is there a neighborhood frequency effect in English? Evidence from reading and lexical decision

What is the effect of a word's higher frequency neighbors on its identification time? According to activation-based models of word identification (J. Grainger & A. M. Jacobs, 1996; J. L. McClelland & D. E. Rumelhart, 1981), words with higher frequency neighbors will be processed more slowly than words without higher frequency neighbors because of the lexical competition mechanism embodied in these models. Although a critical prediction of these models, this inhibitory neighborhood frequency effect has been elusive in studies that have used English stimuli. In the present experiments, the effect of higher frequency neighbors was examined in the lexical decision task and when participants were reading sentences while their eye movements were monitored. Results suggest that higher frequency neighbors have little, if any, effect on the identification of English words. The implications for activation-based models of word identification are discussed.     

Morphological knowledge as revealed in children's spelling accuracy and reports of spelling strategies

The goal of the current research was to assess whether children can make strategic use of morphological relations among words to spell. French-speaking children in Grade 4 spelled three word types: (a) phonological words that had regular phoneme-grapheme correspondences, (b) morphological words that had silent consonant endings for which a derivative revealed the silent ending, and (c) lexical words that had silent consonant endings for which no familiar derivative revealed the ending. Children were also asked to provide immediate retrospective reports of the strategies used to spell each word. Two experiments (Ns = 46 and 39) were conducted. As expected, children in Grade 4 spelled phonological words more accurately than they did words with silent consonant endings. In addition, children spelled morphological words more accurately than they did lexical words. Reports of using retrieval were associated with accurate performance across word types. Importantly, reports of using morphological strategies to spell morphological words were associated with a similar level of accuracy, as were reports of using retrieval. Even though children reported using a phonological strategy frequently across all word types, this strategy was associated with accurate performance only for spelling phonological words. Experiment 2 replicated the results of Experiment 1 with another set of stimuli and also showed that children's morphological awareness predicted their spelling accuracy for morphological words as well as the reported frequency of morphological strategy use. In sum, the findings revealed that most children showed evidence of adaptive strategy use.     

Morphological priming effect: the role of surface frequency.

Two cross-modal experiments were conducted to investigate the format of lexical representation of suffixed derived words and their stems. The results show that only low frequency suffixed words (as opposed to high frequency suffixed words) yield a full priming effect of their stems. By contrast, a stem (e.g., travail) does not fully prime words belonging to the same morphological family (e.g., travailleur), although it primes high frequency suffixed words more than it does low frequency words. To account for these findings we propose a model in which the stem and high frequency affixed words are represented both as full forms and as decomposed morphemes while low frequency affixed words are represented only in a decomposed form.     

Decomposition: to what extent? The case of Turkish.

It has been proposed that in agglutinative languages, lexical access of morphologically complex words must involve decomposition rather than full listing (Frauenfelder & Schreuder, 1992; Hankamer, 1989). We tested this proposal in Turkish using a simple lexical decision task. Results show that multimorphemic words that consist of frequent affixes are processed as fast as monomorphemic words. This finding suggests that in languages with rich morphology, not all multimorphemic words are accessed in a decomposed form. To the extent that morphemes are in frequent use, they may induce whole-word rather than decompositional lexical access.     

Linking elements in Dutch noun-noun compounds: constituent families as analogical predictors for response latencies

This study addresses the choice of linking elements in novel Dutch noun-noun compounds. Previous off-line experiments (Krott, Baayen, & Schreuder, 2001) revealed that this choice can be predicted analogically on the basis of the distribution of linking elements in the left and right constituent families, i.e., the set of existing compounds that share the left (or right) constituent with the target compound. The present study replicates the observed graded analogical effects under time pressure, using an on-line decision task. Furthermore, the analogical support of the left constituent family predicts response latencies. We present an implemented interactive activation network model that accounts for the experimental data.     

Sublexical and morphological information in speech processing

Three experiments investigated the impact of syllabic boundary information and of morphological structure on performance in a sequence-monitoring task. In sequence monitoring, participants detect pre-specified sequences of phonemes in spoken carrier words. Sequences corresponded to the first syllable of the carrier word, to its first morpheme, or simultaneously to both. The data from Experiments 1 and 2, using different variants of the monitoring task, showed a strong impact of syllable boundary cues on monitoring latencies. An effect of morphological match between targets and carrier words was also evident. Experiment 3, in which parts of the spoken carrier words were cross-spliced, revealed that syllabic boundary information takes precedence over morphological information. The results are in line with an early process of speech processing, in which syllabic cues are used to aid lexical access. The morphological effect is better understood as a later, probably lexical, contribution of morphological decomposition to monitoring performance.     

Neuromagnetic evidence for the timing of lexical activation: an MEG component sensitive to phonotactic probability but not to neighborhood density

Evidence from electrophysiological measures such as ERPs (event-related potentials) and MEG (magnetoencephalography) suggest that the first evoked brain response component sensitive to stimulus properties affecting reaction times in word recognition tasks occurs at 300-400 ms. The present study used the stimulus manipulation of Vitevich and Luce (1999) to investigate whether the M350, an MEG response component peaking at 300-400 ms, reflects lexical or postlexical processing. Stimuli were simultaneously varied in phonotactic probability, which facilitates lexical activation, and in phonological neighborhood density, which inhibits the lexical decision process. The present results indicate that the M350 shows facilitation by phonotactic probability rather than inhibition by neighborhood density. Thus the M350 cannot be a postlexical component.     

Processing reduced word forms: the suffix restoration effect

Listeners cannot recognize highly reduced word forms in isolation, but they can do so when these forms are presented in context (Ernestus, Baayen, & Schreuder, 2002). This suggests that not all possible surface forms of words have equal status in the mental lexicon. The present study shows that the reduced forms are linked to the canonical representations in the mental lexicon, and that these latter representations induce reconstruction processes. Listeners restore suffixes that are partly or completely missing in reduced word forms. A series of phoneme-monitoring experiments reveals the nature of this restoration: the basis for suffix restoration is mainly phonological in nature, but orthography has an influence as well.     

The cumulative semantic cost does not reflect lexical selection by competition

The cumulative semantic cost describes a phenomenon in which picture naming latencies increase monotonically with each additional within-category item that is named in a sequence of pictures. Here we test whether the cumulative semantic cost requires the assumption of lexical selection by competition. In Experiment 1 participants named a sequence of pictures, while in Experiment 2 participants named words instead of pictures, preceded by a gender marked determiner. We replicate the basic cumulative semantic cost with pictures (Exp. 1) and show that there is no cumulative semantic cost for word targets (Exp. 2). This pattern was replicated in Experiment 3 in which pictures and words were named along with their gender marked definite determiner, and were intermingled within the same experimental design. In addition, Experiment 3 showed that while picture naming induces a cumulative semantic cost for subsequently named words, word naming does not induce a cumulative semantic cost for subsequently named pictures. These findings suggest that the cumulative semantic cost arises prior to lexical selection and that the effect arises due to incremental changes to the connection weights between semantic and lexical representations.