Interaction of Frequency and Inflectional Status: An Approach From Discriminative Learning

High frequency of occurrences has been associated with phonetic reduction, on one hand, and phonetic enhancement, on the other hand. This study first looks into the possibility that these opposite frequency effects are at least partially due to the different inflectional status of the items being investigated. Based on tongue position data from a spontaneous speech corpus of German, we found that stem vowels in inflected words tended to be hyper-articulated (i.e., phonetic enhancement), whereas those in non-inflected words tended to be articulated with more centralized tongue positions (i.e., phonetic reduction). This observed modulation by inflectional status is subsequently investigated from the perspective of distributional semantics. Using the discriminative lexicon model to study relations between word embeddings and word forms, we observed that word-final triphones of inflected words received more support from their embeddings (i.e., meanings) compared with non-inflected words. Furthermore, replacement of the two-level factorial predictor of inflectional status with the amount of semantic support led to substantial improvement in model fit. These results suggest direct relationships between word forms and meanings and argue for further investigation of form-meaning isomorphies and their consequences for cognitive models of speech production.

Authors

Institutions

Publication Details

Journal
Language and Speech
Published
2026-08-25
DOI
https://doi.org/10.1177/00238309261474056
Primary Topic
Neural Networks and Applications
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Interaction of Frequency and Inflectional Status: An Approach From Discriminative Learning

Fabian Tomaschek, Motoki Saito, R. Harald Baayen
Language and Speech
Neural Networks and Applications
article

Interaction of Frequency and Inflectional Status: An Approach From Discriminative Learning

Fabian Tomaschek, Motoki Saito, R. Harald Baayen
article en

Abstract

High frequency of occurrences has been associated with phonetic reduction, on one hand, and phonetic enhancement, on the other hand. This study first looks into the possibility that these opposite frequency effects are at least partially due to the different inflectional status of the items being investigated. Based on tongue position data from a spontaneous speech corpus of German, we found that stem vowels in inflected words tended to be hyper-articulated (i.e., phonetic enhancement), whereas those in non-inflected words tended to be articulated with more centralized tongue positions (i.e., phonetic reduction). This observed modulation by inflectional status is subsequently investigated from the perspective of distributional semantics. Using the discriminative lexicon model to study relations between word embeddings and word forms, we observed that word-final triphones of inflected words received more support from their embeddings (i.e., meanings) compared with non-inflected words. Furthermore, replacement of the two-level factorial predictor of inflectional status with the amount of semantic support led to substantial improvement in model fit. These results suggest direct relationships between word forms and meanings and argue for further investigation of form-meaning isomorphies and their consequences for cognitive models of speech production.

Language and Speech
University of Bern (CH), Carl von Ossietzky Universität Oldenburg (DE), University of Tübingen (DE)
Openalex Percentile: Top 99%
Neural Networks and Applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.