The writers begin by briefly reviewing a morphemic level of word representation rather than a lexical, which I believe refers to full-word storage. These two versions of representation are linked when the theory is broken into two categories: prelexical and supralexical representation. Prelexical hypothesizes that a word is broken into its morphemes before being fully activated, whereas supralexical suggests that the word is first seen in its full, lexical form, and then the morphemes are activated.
Just as words that are in similar contextual categories (such as boat and dock) prime each other, words and morphemes are primers. An example used is â??departure,â?? which primes â??depart.â?? However, in the case of â??brother,â?? it does not prime for â??brothâ?? because it is different both phonetically and semantically.
The article then moves on to French, which gives every noun a gender. In 89.5% of the cases, the gender is arbitrary (for instance, the word â??girlâ?? would not be arbitrary because the noun is inherently gendered). The authorsâ?? experiment-fueling question revolves around how gender is lexically stored and recalled. When there are clues to the gender, when they follow rules, the gender is more easily accessible than when the gender is an exception to a rule or strictly based on memorization.
One of the topics we discussed in class was whether our lexicons are stored based upon rules or upon individual words. Although we couldnâ??t decideâ??no surprise, neither can scientistsâ??this seems to a similar question applied to gendered words in French. Is each word stored separately with its own gender, or do we process each based upon the rules that we know? Of course most wordsâ??those without constructions that point one way or anotherâ??are exceptions to any of the rules and must just be memorized. The idea that those that follow a pattern may be processed faster seems to favor the idea that we function better with guiding rules.
Although I wonâ??t get into the details of the experiments, the results are important. As far as I understand, the first experiment showed that as a word is broken into its morphemic parts, its gender is also activated. Both experiments demonstrated, as earlier hypothesized, that when the gender is congruent with a pattern from the rest of the language, the gender is accessed quicker. The second experiment further supported the firstâ??s findings that a wordâ??s gender is accessed along with its morphemic parts.
The results of the three experiments are best summarized in Fanny Meunier, Alix Seigneuric, and Elsa Spinelliâ??s own words, which, at least, will ensure that I donâ??t misrepresent their findings. They write, â??Overall our results clearly show that during gender identification, nouns made of morpheme-like units are decomposed. When the units carry conflicting gender information, gender decisions are delayed. Our results also allow to define more precisely the nature of the information chunk that is extracted and that have been previously called â??endingsâ??: morphological suffixes.â??