Andrew Wedel

Exemplar models, evolution and language change

Abstract Evidence supporting a rich memory for associations suggests that people can store perceptual details in the form of exemplars. The resulting particulate model of category contents allows the application of evolution theory in modeling category change, because variation in categorized percepts is reflected in the distribution of exemplars in a category. Within a production-perception feedback loop, variation within an exemplar-based category provides a reserve of variants that can serve as the seeds for shifts in the system over time through random or selection-driven asymmetries in production and perception. Here, three potential pathways for evolutionary change are identified in linguistic categories: pruning of lines of inheritance, blending inheritance and natural selection. Simulations of each of these pathways are shown within a simple exemplar-based model of category production and perception, showing how consideration of evolutionary processes may contribute to our understanding of linguistic category change over time.

Feedback and regularity in the lexicon

Phonologies are characterized by striking regularity, from the stereotyped phonetic characteristics of particular allophones to the contextually conditioned alternations between them. Most models of grammar account for this regularity by hypothesizing that there is only a limited set of abstract symbols available for expressing underlying forms, and that an independent grammar algorithm predictably transforms these abstract symbol sequences into an output representation. However, this explanation for regularity is called into question by much recent research suggesting that the mental lexicon records rich phonetic detail that directly informs production. Given evidence for persistent biases favoring previously experienced forms at a number of levels of production and perception, I argue here that positive feedback within a richly detailed lexicon can serve as a source of regularity over many cycles of production and perception. Using simulation as a tool, I show that under the influence of positive feedback, gradient biases in usage can convert an initially gradient and variable distribution of lexical behaviors into a more categorical, and often simpler, pattern.

High functional load inhibits phonological contrast loss: a corpus study.

For nearly a century, linguists have suggested that diachronic merger is less likely between phonemes with a high functional load--that is, phonemes that distinguish many words in the language in question. However, limitations in data and computational power have made assessing this hypothesis difficult. Here we present the first larger-scale study of the functional load hypothesis, using data from sound changes in a diverse set of languages. Our results support the functional load hypothesis: phoneme pairs undergoing merger distinguish significantly fewer minimal pairs in the lexicon than unmerged phoneme pairs. Furthermore, we show that higher phoneme probability is positively correlated with merger, but that this effect is stronger for phonemes that distinguish no minimal pairs. Finally, within our dataset we find that minimal pair count and phoneme probability better predict merger than change in system entropy at the lexical or phoneme level.

Functional Load and the Lexicon: Evidence that Syntactic Category and Frequency Relationships in Minimal Lemma Pairs Predict the Loss of Phoneme contrasts in Language Change

All languages use individually meaningless, contrastive categories in combination to create distinct words. Despite their central role in communication, these “phoneme” contrasts can be lost over the course of language change. The century-old functional load hypothesis proposes that loss of a phoneme contrast will be inhibited in relation to the work that it does in distinguishing words. In a previous work we showed for the first time that a simple measure of functional load does significantly predict patterns of contrast loss within a diverse set of languages: the more minimal word pairs that a phoneme contrast distinguishes, the less likely those phonemes are to have merged over the course of language change. Here, we examine several lexical properties that are predicted to influence the uncertainty between word pairs in usage. We present evidence that (a) the lemma rather than surface-form count of minimal pairs is more predictive of merger; (b) the count of minimal lemma pairs that share a syntactic category is a stronger predictor of merger than the count of those with divergent syntactic categories, and (c) that the count of minimal lemma pairs with members of similar frequency is a stronger predictor of merger than that of those with more divergent frequencies. These findings support the broad hypothesis that properties of individual utterances influence long-term language change, and are consistent with findings suggesting that phonetic cues are modulated in response to lexical uncertainty within utterances.

Lexical contrast maintenance and the organization of sublexical contrast systems

Abstract Variationist/evolutionary models of phonology assume a causal chain that links biases at the utterance level to the development and consolidation of abstract phonological patterns over time. Some of the properties of linguistic cognition that have been proposed to underlie this chain are (i) storage of experienced detail at multiple levels of description, (ii) feedback between perception and production, (iii) a similarity bias in the production and perception of variation, and (iv) enhancement of cues to potentially ambiguous lexical items in usage. I review evidence for these properties and argue that they interact to provide a pathway for individual usage events to influence the evolution of contrastive sublexcal category systems, i.e phoneme inventories. Specifically, the proposed Network-Feedback model predicts that the organization of sublexical category systems is shaped by a conflict between a general drive toward greater similarity among sublexical categories on the one hand, and a bias toward maintaining contrast between tokens of competing lexical categories on the other. The model provides testable hypotheses about the conditions favoring phoneme merger, chain-shifts, and phonemic splits, and more generally about the influence of lexical contrast on the packing of sublexical categories along gestural/perceptual dimensions. Finally, this pathway of change is consistent with proposals that sublexical categories such as features and segments are not primitives of language, but emerge through more general properties of performance, perception, categorization and learning.

Category competition drives contrast maintenance within an exemplar-based production/perception loop

The evolution of competing lexical categories is simulated within a model in which lexical outputs are organized as sequences of articulatory gestures. When exemplar-based categories compete for assignment and storage of incoming exemplars in a production/storage loop, contrast between categories spontaneously emerges and remains stable, driven by the differences in storage consistency between more contrastive and less contrastive variants. Further, when lexical outputs are biased toward use of previously produced gestures, the set of exemplars in the lexicon evolve to be derived from a small set of contrastive units used in combination, despite the absence of direct selection for contrast at the sub-lexical level.

Auditory masked priming in Maltese spoken word recognition

This study investigated lexical access in Maltese, an understudied Semitic language. We report here on a series of four lexical decision experiments designed to test the hypothesis that the consonantal root and the word pattern may each prime lexical access in Maltese. Priming of morphologically related forms is generally taken as evidence consistent with morphological decomposition in processing. Here, we used two speech priming techniques: auditory priming in which primes and targets were equally audible, and auditory masked priming in which primes are masked from conscious perception by volume-attenuation and compression. Our results show priming of targets by forms sharing a consonantal root, but not by forms sharing a word pattern.

The phonetic specificity of competition: Contrastive hyperarticulation of voice onset time in conversational English

Abstract Competition between words in the lexicon is associated with hyperarticulation of phonetic properties in production. This correlation has been reported for metrics of competition varying in the phonetic specificity of the relationship between target and competitor (e.g., neighborhood density, onset competition, cue-specific minimal pairs). Sampling a systematic array of competition metrics, we tested their ability to predict voice onset times in both voiced and voiceless word-initial stops of conversational English. Linear mixed effects models were compared according to their corrected Akaike’s Information Criterion (AICc) values. High-performing models were evaluated using evidence ratios, with the competition metrics of top-performing models tested for significance using nested model comparisons. Words with a minimal pair defined for initial stop voicing were contrastively hyperarticulated, with shorter voice onset times for voiced stops and longer voice onset times for voiceless stops. No other competition metric reliably predicted hyperarticulation for both stop types. These results suggest that contrastive hyperarticulation is phonetically specific, increasing the perceptual distance between target and competitor.

Fishing the best pool for novel ribozymes

Novel RNA enzymes, or ribozymes, are sought in large pools of random RNA sequences. Because of the large number of random positions in an individual pool molecule, only a vanishingly small fraction of the possible sequences are actually present. Even so, increasing the length of the individual pool molecules significantly increases the probability of finding a particular complex ribozyme. Because ribozymes are typically composed of conserved sequences interleaved with regions that can vary in sequence and length, a longer molecule allows a greater number of possible arrangements of a given ribozyme motif, increasing the likelihood that it will be present in the pool. Once a ribozyme motif has been found, rational and irrational optimization techniques can be used to identify related ribozyme sequences with greater activity.

The Co-evolution of Speech and the Lexicon: The Interaction of Functional Pressures, Redundancy, and Category Variation.

The sound system of a language must be able to support a perceptual contrast between different words in order to signal communicatively relevant meaning distinctions. In this paper, we use a simple agent-based exemplar model in which the evolution of sound-category systems is understood as a co-evolutionary process, where the range of variation within sound categories is constrained by functional pressure to keep different words perceptually distinct. We show that this model can reproduce several observed effects on the range of sound variation. We argue that phonological systems can be seen as finding a relative optimum of variation: Efficient communication is sustained while at the same time, hidden category variation provides pathways for future evolution.