ChangeLing Lab

Language Change and Empirical Linguistics at CMU

new-logo.png

5407 Gates Hillman Complex

Language Technologies Institute

Carnegie Mellon University

ChangeLing Lab is Carnegie Mellon University’s only research lab focused on understanding how languages change, and how these patterns of change shape the way that languages are at any given point in time, from a computational perspective. We are interested in phonetics, phonology, and morphology (whether diachronic or synchronic), emergent communication, and have a special concern for the use of language science to benefit people with disabilities.

ChangeLing is lead by David R. Mortensen, an Associate Research Professor in the Language Technologies Institute. It consists, additionally, of graduate students, former LTI students who still collaborate with David, and visitors.

If you are interested in joining ChangeLing, please email David at dmortens@cs.cmu.edu with a CV and a description of what work you would like to do with us. Please take the following into account:

  • We are only concerned with work that has some linguistic angle (either it uses linguistics or it is useful for linguists). Students who are concerned with machine learning for its own sake would be better served by another lab.
  • We are interested in large language models, but only with respect to their language and linguistic reasoning capabilities. Our lab is not a good place to do general, engineering-focused or fundamental research on LLMs.

News

Sep 09, 2026 We are proud to announce that Phone Segmentation and Recognition through Phonological Activation Mapping, or SPAM (S3M-based Phonological Activation Mapping) for short, has been accepted to SLT 2026!
Phonetic structure is already present in the representations of self-supervised speech models (S3Ms), and we show that they can be steered to solve phone segmentation and recognition at the same time. The pipeline is simple and intuitive: S3M representations, phonological feature activations, and two gradient-descent-free prediction heads. Our method requires less than a minute of phonetic transcriptions and generalizes well to phones unseen during training!
Apr 07, 2026 The papers
  • POWSM: A Phonetic Open Whisper-Style Speech Foundation Model (main, top 5%)
  • PRiSM: Benchmarking Phone Realization in Speech Models (main)
  • Communicating in Emergent Language with an Induced Morphological Phrasebook (main)
  • [b] = [d] - [t] + [p]: Self-supervised Speech Models Discover Phonological Vector Arithmetic (findings)
  • Linear Script Representations in Speech Foundation Models Enable Zero-Shot Transliteration (findings)
  • PBEBench: A Multi-Step Programming by Examples Reasoning Benchmark inspired by Historical Linguistics (findings)
were accepted to ACL 2026.
Oct 12, 2025 David will give the Colloquium talk at LTI, “The Reconstruction Will Not Be Supervised.”

Latest Posts