Lexical flexibility in English: A preliminary study

Today I gave a talk to the linguistics department at the College of William & Mary titled 'Lexical flexibility in English: A preliminary study', in which I presented initial results from my dissertation work, focusing on my English data. Check out the abstract below, or download the slides here!


Parts of speech such as noun, verb, and adjective were once thought to be clearly defined and universal to all languages. Today, the definition, identification, and crosslinguistic comparison of parts of speech is one of the most controversial topics in linguistic typology (the study of crosslinguistic patterns). Since the emergence of cognitive linguistics in the last several decades, linguists have generally accepted that parts of speech are fuzzy and prototypal (clustered around prototypes) rather than categorical and clearly distinguished. Moreover, some linguists claim that certain languages fail to distinguish particular parts of speech at all: Quechua is said to lack adjectives; in Nahuatl every word is said to be capable of functioning like a verb; Nuuchahnulth is thought to lack a distinction between nouns and verbs. In the extreme case, Riau Indonesian is claimed to lack any parts of speech entirely.

As a result, some linguists in recent decades suggest that words in some languages are not associated with any particular part of speech. Words in these languages are said to be flexible, capable of functioning as different parts of speech depending on context, without the need for derivational morphology. Other linguists deny the existence of such flexible words, stating that the appearance of lexical flexibility is the result of zero-marked derivation or conversionfrom one part of speech to another, resulting in two separate, homophonous words. Very few studies treat these possible cases of flexibility as an object of study in themselves. We are left with the question whether the empirical data on potentially flexible words support one interpretation or the other. Put differently, just how fuzzy and flexible are parts of speech in the first place?

This talk presents preliminary results from a quantitative study of lexical flexibility in English (which is often claimed to be a fairly flexible language) designed to take a first step towards answering these questions. Using the spoken portion of the Open American National Corpus (OANC), I manually annotated over 25,000 words of English for their function—nominal, verbal, or adjectival—and then determined how flexible each word is by quantifying how often that word is used for different parts of speech. Preliminary results suggest that English words vary drastically in how flexible they are, but that most English words exhibit flexibility to at least a small degree. Though additional data are needed, initial results also suggest that certain semantic domains—specifically body part terms—are more likely to exhibit lexical flexibility than others. On the whole, however, both quantitative and qualitative analysis of the data seem to support the position that English is not an especially flexible language.