A syntax-lexicon trade-off in language production

Abstract

Spoken language production involves selecting and assembling words and syntactic structures to convey one's message. Here we probe this process by analyzing natural language productions of individuals with primary progressive aphasia (PPA) and healthy individuals. Based on prior neuropsychological observations, we hypothesize that patients who have difficulty producing complex syntax might choose semantically richer words to make their meaning clear, whereas patients with lexicosemantic deficits may choose more complex syntax. To evaluate this hypothesis, we first introduce a frequency-based method for characterizing the syntactic complexity of naturally produced utterances. We then show that lexical and syntactic complexity, as measured by their frequencies, are negatively correlated in a large (n = 79) PPA population. We then show that this syntax-lexicon trade-off is also present in the utterances of healthy speakers (n = 99) taking part in a picture description task, suggesting that it may be a general property of the process by which humans turn thoughts into speech.

Keywords: primary progressive aphasia; syntactic complexity; syntax frequency; syntax–lexicon trade-off; word frequency.

Fig. 1.

Extracting syntactic rules of a dependency parse. (A) Sentences produced by participants were annotated and disfluencies removed. (B) Sentences were parsed using the Stanford Lexicalized Parser Package. Dependency heads, underlined, were identified by the parser. (C) The combinations of heads and dependents, preserving the order in which they appeared in the original sentence, were extracted as syntactic rules. This sentence resulted in three syntactic rules. The analyses we present here are by utterance (sentence). Thus, for this utterance, we would take the average syntax frequency of the three syntactic rules and the average word frequency for the content words in this sentence.

Fig. 2.

(A) The density graphs of content word frequency and (B) syntax frequency of utterances. (C) The box plots of content word frequency per individual in each group. (D) The box plots of syntax frequency per individual in each group. The box plots show the median, 25th, and 75th percentiles of the data. Error bars represent 95% CI.

Fig. 3.

The proportions of use of the 20 most common syntactic rules of the Switchboard by PPA patients and healthy controls. The last set of bars shows all other lower-frequency rules.

Fig. 4.

Syntax–lexicon scatterplots. (A) Utterance level in PPA patients (n = 79) and healthy controls (n = 53). (B) Subject level in PPA patients (dashed line) and healthy controls (solid line represents all data points). (C) Utterance level in healthy individuals (n = 99). (D) Subject level in healthy individuals.