There is a correlation between the phonological shape of a word and the word’s probability in use. Less probable words tend to be longer and more probable words shorter (see Piantadosi et al. , Zipf ). This has been attributed to the lexicon evolving for efficient communication (Zipf ). To identify less probable words, listeners need more information from the segments in the phonological word itself. In this case, longer lengths for less probable words mean a greater amount of information to be used in word identification. However, this does not take into account how listeners actually process words. Research in spoken word recognition has shown that words are processed incrementally and some segments may in fact be more informative (Allopenna et al. , Luce and Pisoni , van Son and Pols , Weber and Scharenborg ). Here, we use corpus data from American English to provide evidence that less probable words contain more informative segments. We also show that the distribution of segmental information is correlated with the word’s probability and that less probable words contain more of their total information in the early segments. We discuss these findings and possible evolutionary avenues for language to reach this state. This work provides support for the idea that the words in the lexicon evolve under pressure for efficient communication.

A canonically agglutinative language or morphological pattern is traditionally analyzed as building words out of independent morphemes. Using data from Choguita Rarámuri (Uto-Aztecan), we attempt to quantify this notion by examining the extent to which meanings are predictable from their exponents without reference to context. We show that two-layer connectionist networks, computational models that map form onto meaning directly, can be used for this purpose. We also show that learning the meanings of morphemes can pose significant challenges to such models and constrains the design of the learning algorithm. In particular, models trained to equilibrium tend to focus on unreliable cues to the meanings they try to predict, especially when trained on a small corpus typical of underresourced languages. Some of these issues can be alleviated by a slow learning rate. However, one issue — which we call the problem of spurious excitement — is shown to be inherent to the learning algorithm, and always arises by the time the model achieves equilibrium. Spurious excitement means that a cue becomes associated with a meaning that it does not co-occur with, simply because of co-occurring with cues that disfavor the meaning. This case raises larger implications with respect to the type of learning mechanism involved in the acquisition of natural languages. Solutions to spurious excitement are discussed. The logistic activation function is shown to improve the performance of the model in detecting reliable cues to meanings that recur across many word types (i.e., cues of high type frequency), as well as eliminating spurious excitement.