Morphological decomposition: more than islands of regularity?

A wealth of psycholinguistic evidence has shown that words, before being visually recognized, decompose into smaller units which seem to correspond to morphemes (Rastle et al., 2000, 2004). Such a procedure of morphological decomposition seems to occur in words that are made of more than one morpheme, independently of whether they are semantically transparent (e.g., driver, which means “someone who drives”) or not (e.g., brother, which does not mean “someone who broths”). Decomposition, however, does not seem to occur in words that contain a root plus an additional, non-morphemic string (e.g., brothel, where el is not an English suffix). Current models of visual word processing assume the morpho-orthographic decomposition hypothesis, whereby decomposition obligatorily relies on islands of regularity – namely, statistical orthographic regularities across letter strings, so that a string such as er is detected as a single morpho-orthographic unit, but el is not (Rastle & Davis 2008). As such, decomposition is also argued to occur at early stages of word processing and before contact with the lexicon — namely, before semantic information is accessed, so that decomposition occurs regardless of semantic transparency. In my doctoral dissertation, I built on these findings and use the visual masked priming response to test the sensitivity of decomposition to any of the following higher-level linguistic properties (in addition to islands of regularities): phono-orthographic syllabification, whole-word frequency, phonological conditioning, and syntactic well-formedness.

  1. SYLLABIFICATION. The hypothesis that decomposition goes through a mechanism of low-level syllabic chunking was first tested because it may explain why words that contain a root but not a second morpheme do not decompose (brothel↛ broth-el; ‘brothel effect’; Rastle et al., 2004), but non-words that contain a root but not a second morpheme do (flexire→flex-ire; ‘flexire effect’; Morris et al., 2011). One possibility is that the flexire effect results from the syllabic chunking operation (flexire → FLEX.IRE → flex-ire; where the dot ‘.’ signals a loose syllabic boundary). The brothel effect would then be a consequence of averaging the priming response to words with a syllabic, non-affixal ending (brothel → BROTH.EL → broth-el) together with the non-priming response to words with a non-syllabic, non-affixal ending (i.e., against → AGAINST ↛ again-st). In our experiments, masked priming was elicited in two different orthographically related, but morphologically unrelated conditions: a syllabic condition, in which the prime word contained the corresponding target word, plus a syllabic unit (e.g., ban.jo-BAN, where the dot “.” signals the syllabic boundary); and a non-syllabic condition, in which the prime word contained the corresponding target word, plus a consonantal unit (e.g., starch-STAR, where the underlined string signals the extra consonantal unit). Results showed that both conditions did not elicit priming, thus suggesting that syllabification does not drive morphological decomposition (at least in the visual modality). These results potentially imply that decomposition is affected by the lexicality of the stimulus: in words, decomposition occurs only if they can be transparently broken down into morphemes; in non-words, decomposition occurs as long as they contain at least one morpheme.

  2. DOMINANCE. The term ‘dominance’ refers to the asymmetrical ratio between frequencies of the forms in the same nominal paradigm: for example, a word such as windows is plural-dominant because the surface frequency of the plural form (windows) is higher than the surface frequency of the singular form (window); a word such as worlds is instead singular-dominant, because the surface frequency of the plural form (worlds) is lower than the surface frequency of the singular form (world). Lexical decision studies have reported that plural-dominant plural forms are recognized faster than singular-dominant plural forms. The hypothesis that dominance potentially impinges on decomposition was tested to understand whether word frequency is accessed during decomposition, or, rather, at later stages of processing. Masked priming was elicited in two different morphological conditions: in the sgdom-sgdom condition, the pairs contained singular-dominant plural forms as both primes and targets (e.g., worlds-HEAVENS); in the pldom-sgdom condition, the pairs contained a plural-dominant plural prime and a singular-dominant plural target (e.g., windows-GODS). Results showed that priming arose in the sgdom-sgdom condition, but not in the pldom-sgdom condition, thus suggesting that dominance asymmetries (and therefore word frequency) impinge on decomposition.

  3. SYNTACTIC WELL-FORMEDNESS. The term ‘syntactic well-formedness’ refers to the syntactic categories affixes may and may not attach to; for example, the affix –(i)ble attaches to verbs, but not to nouns: e.g., flex-ible, but bliss-able. The hypothesis that decomposition is affected by violations of syntactic well-formedness restrictions was therefore tested to explore whether decomposition accesses syntactic properties of morphemes. In our experiments, the masked priming response was elicited for both syntactically well-formed and ill-formed prime stimuli – e.g., purity-PURE vs skittity-SKIT. These results suggest that decomposition is not affected by syntactic well-formedness.

Taken together, these results suggest that while not being affected by syntactic well-formedness, decomposition is indeed sensitive to whole-word lexicality and whole-word frequency. This clear-cut divide between properties that may and may not impinge on decomposition might be due to the fact that both syntactic processing (i.e., the process whereby syntactic categories and affixal restrictions are retrieved and checked) and semantic processing (i.e., the process whereby whole-word meaning is accessed) require at least two mechanisms to occur: (i) recombination of decomposed morphemes (along the lines of Taft 2004), and, eventually, (ii) access to compatible syntactic and semantic representations. Given that these are fairly complex, these representations indeed take more time to access fully, which, therefore, may not occur until after decomposition.

At the face value, the fact that whole-word lexicality and frequency may impinge on early decomposition procedures is surprising. These two properties are generally considered to be accessed at late stages of processing and their potential implications onto decomposition seems logically contradictory, under the current theories of early decomposition: how can an early process be affected by properties that concern the whole stimulus and, as such, are accessed only at later stages? A model of early visual decomposition is proposed that is able to account for these findings without necessarily challenging the morpho-orthographic decomposition hypothesis. In this model, decomposition accesses whole-word lexicality and frequency through a multi-step mechanism that first generates multiple possible morpho-orthographic decomposition patterns of the visual stimulus and then evaluates them in parallel in order to choose the optimal candidate for activation.

The investigation above has inspired a number of follow-up studies on morphological decomposition. Click on the links below for more information!