--- title: "Improving Automated Prediction of English Lexical Blends" description: "Lexical blends such as combinations of two words into one (keeping the start of the… 74% exact match. 13 references, cited by 3 papers here." authors: ["Jarem Saunders"] year: 2023 venue: "Proceedings of the 20th SIGMORPHON workshop on Computational Research in Phonetics, Phonology, and Morphology" source: "https://aclanthology.org/2023.sigmorphon-1.10/" doi: "10.18653/v1/2023.sigmorphon-1.10" licence: "CC BY 4.0 (ACL Anthology)" licence_url: "https://creativecommons.org/licenses/by/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/saunders-2023-improving-automated-prediction-english-lexical/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/saunders-2023-improving-automated-prediction-english-lexical/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2023](https://md.research.elamur.ai/free/en/papers/2023/page/13/) / Improving Automated Prediction of English Lexical Blends Through the Use of Observable Linguistic Features # Improving Automated Prediction of English Lexical Blends Through the Use of Observable Linguistic Features **Authors:** Jarem Saunders · **Year:** 2023 · **Venue:** [Proceedings of the 20th SIGMORPHON workshop on Computational Research in Phonetics, Phonology, and Morphology](https://md.research.elamur.ai/free/en/venues/sigmorphon/) ([SIGMORPHON 2023](https://md.research.elamur.ai/free/en/venues/sigmorphon/2023/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2023.sigmorphon-1.10/) · [PDF](https://aclanthology.org/2023.sigmorphon-1.10.pdf) · [DOI 10.18653/v1/2023.sigmorphon-1.10](https://doi.org/10.18653/v1/2023.sigmorphon-1.10) · [OpenAlex](https://openalex.org/W4385718117) **Citations:** 0 citations (OpenAlex, 2026-10-14) · 3 in our library, of the 149,439 papers whose bibliographies we read ## Summary Lexical blends such as combinations of two words into one (keeping the start of the first and the end of the second) are notoriously hard to predict. Earlier computational attempts, a multi-tape FST by Deri & Knight (2015) and a character-embedding LSTM by Gangal et al. (2017), got fewer than half of blends right. This paper takes a different route. It enumerates every candidate that joins a prefix of word 1 with a suffix of word 2, cutting only at grapheme/phoneme alignment points. Each candidate is described by interpretable phonological features based on Arndt-Lappe & Plag (2013), plus a BLICK phonotactic score. Simple learners (LASSO, second-order Polynomial regression, Random Forest) assign each candidate a probability, and the highest-scoring candidate is chosen. Across three corpora, the Polynomial model with the full feature set clearly beats the previous benchmarks, reaching up to 74% exact matches. The work shows that explicit, observable linguistic features can outperform latent-space neural models on rare, irregular word-formation processes. ## Key points - Data came from three corpora, filtered to blends that keep the start of word 1 and the end of word 2: Deri & Knight (322 blends), Gangal et al. (1092), and Shaw (2014), curated from Thurner's 1993 dictionary (1096). All trials used 10-fold cross-validation. - Features cover: - syllable lengths and medial overlap; - distances from word edges to primary stress and to the switchpoint; - whether the switchpoint falls on an onset, nucleus or coda boundary; - preservation of stress, segments and syllables; - the BLICK markedness score, a MaxEnt grammar trained on the CMU pronouncing dictionary. - Results for the best learner (Polynomial, full features), measured by exact match and Levenshtein edit distance: | Corpus | This paper | Previous benchmark | |---|---|---| | Deri & Knight | 64.42% (edit distance 0.72) | 45.39% (edit distance 1.59) | | Gangal et al. | 59.51% | 48.75% | | Shaw | 74.13% | — | - Only LASSO failed to beat the LSTM benchmark on the Gangal corpus. The full feature set beat a hand-pruned, low-collinearity subset, even though the full set is highly collinear. - Error analysis of 100 wrong outputs showed a bias toward keeping too much material. Word 1 was over-preserved 40 times versus cut too short 15 times; for word 2 the figures were 33 versus 18. ## Interesting ideas - Instead of decoding a form step by step, enumerate every structurally allowed candidate and let an interpretable scorer rank them; the candidate space itself encodes linguistic constraints. - Candidates whose feature vectors are identical to an earlier candidate are discarded, so the learner never sees indistinguishable positive and negative examples. - A separate phonotactic plausibility score (BLICK) serves as one feature, so the ranker can reject sound sequences that are possible in spelling but un-English. - Errors are split by direction (too much versus too little material kept), which exposes a systematic length bias rather than just a lower accuracy number. - Comparing model outputs with human speakers' brand-new coinages is proposed as truly held-out data, useful even for testing large language models that may have seen existing blends. ## For Mova How these ideas could serve Mova, our small open language model — its files, stages and tests: - **I1** (tokens): Unknown words could be analysed by generating all prefix+suffix splits against known lemmas and ranking them with explicit features, giving an explainable guess at what a novel blend is made of. - **I2** (learn): When training rankers on silver data, removing feature-identical duplicates with conflicting labels would cut noise and make rule induction cleaner. - **I3** (back): When generating names or novel word forms, a phonotactic well-formedness score could filter out unpronounceable outputs. - **I4** (learn): Evaluation reports could classify errors by direction, such as over- versus under-attachment or spans that are too long versus too short, to show which repair rules are needed. - **I5** (learn): Freshly coined test items could act as a contamination-free control when comparing Mova against its LLM teacher tiers. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* ## Institutions - [University of North Carolina at Chapel Hill](https://md.research.elamur.ai/free/en/universities/university-of-north-carolina-at-chapel-hill/) — University, Chapel Hill, United States · 67 works · 16,474 citations 📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works 🧭 Topics (1): [English](https://md.research.elamur.ai/free/en/topics/english/) · 2,263 works ## Cite Cite as: Jarem Saunders. 2023. Improving Automated Prediction of English Lexical Blends Through the Use of Observable Linguistic Features. Proceedings of the 20th SIGMORPHON workshop on Computational Research in Phonetics, Phonology, and Morphology. The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{saunders2023improving, title = {{Improving Automated Prediction of English Lexical Blends Through the Use of Observable Linguistic Features}}, author = {Jarem Saunders}, booktitle = {Proceedings of the 20th SIGMORPHON workshop on Computational Research in Phonetics, Phonology, and Morphology}, year = {2023}, doi = {10.18653/v1/2023.sigmorphon-1.10}, url = {https://doi.org/10.18653/v1/2023.sigmorphon-1.10} } ``` - [← previous GARI: Graph Attention for Relative Isomorphism of Arabic Word Embeddings](https://md.research.elamur.ai/free/en/papers/ali-2023-gari-graph-attention-relative-isomorphism/) - [next → Improving Classroom Dialogue Act Recognition from Limited Labeled Data with Self-Supervised Contrastive Learning Classifiers](https://md.research.elamur.ai/free/en/papers/kumaran-2023-improving-classroom-dialogue-act-recognition/) - [↑ back to papers of 2023](https://md.research.elamur.ai/free/en/papers/2023/page/13/) ## References 13 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Arndt-Lappe, S., & Plag, I. (2013). The role of prosodic structure in the formation of English blends. English Language & Linguistics, 17(3), 537-563. [↗](https://doi.org/10.1017/s1360674313000154) 2. Deri, A., & Knight, K. (2015). How to make a frenemy: Multitape FSTs for portmanteau generation. In Proceedings of the 2015 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (pp. 206-210). [find](https://scholar.google.com/scholar?q=Deri%2C+A%2E%2C+%26+Knight%2C+K%2E+%282015%29%2E+How+to+make+a+frenemy%3A+Multitape+FSTs+for+portmanteau+generation%2E+In+Proceedings+of+the+2015+Conference+of+the+North+American+Chapter+of+the+Association+for+Computationa) 3. Gangal, V., Jhamtani, H., Neubig, G., Hovy, E., & Nyberg, E. 2017. Charmanteau: Character embedding models for portmanteau creation. arXiv preprint arXiv:1707.01176. [↗](https://arxiv.org/abs/1707.01176) 4. Gries, S. T. 2004. Shouldnt it be breakfunch? A quantitative analysis of blend structure in English. [find](https://scholar.google.com/scholar?q=Gries%2C+S%2E+T%2E+2004%2E+Shouldnt+it+be+breakfunch%3F+A+quantitative+analysis+of+blend+structure+in+English%2E) 5. Gries, S. T. 2006. Cognitive determinants of subtractive word formation: A corpus-based perspective. [↗](https://doi.org/10.1515/cog.2006.017) 6. Gries, S. T. 2012. Quantitative corpus data on blend formation: Psycho-and cognitive-linguistic perspectives. Cross-disciplinary perspectives on lexical blending, 252, 145. [↗](https://doi.org/10.1515/9783110289572.145) 7. Hayes, B. 2012. BLICK: a phonotactic probability calculator (manual). [find](https://scholar.google.com/scholar?q=Hayes%2C+B%2E+2012%2E+BLICK%3A+a+phonotactic+probability+calculator+%28manual%29%2E) 8. Hayes, B., & Wilson, C. 2008. A maximum entropy model of phonotactics and phonotactic learning. Linguistic inquiry, 39(3), 379-440. [↗](https://doi.org/10.1162/ling.2008.39.3.379) 9. Kelly, M. H. 1998. To “brunch” or to “brench”: Some aspects of blend structure. [↗](https://doi.org/10.1515/ling.1998.36.3.579) 10. Kubozono, H. 1990. Phonological constraints on blending in English as a case for phonology-morphology interface. Yearbook of morphology, 3, 1-20. [↗](https://doi.org/10.1515/9783112420744-001) 11. Kulkarni, V., & Wang, W. Y. (2018, June). Simple models for word formation in slang. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers) (pp. 1424-1434). [find](https://scholar.google.com/scholar?q=Kulkarni%2C+V%2E%2C+%26+Wang%2C+W%2E+Y%2E+%282018%2C+June%29%2E+Simple+models+for+word+formation+in+slang%2E+In+Proceedings+of+the+2018+Conference+of+the+North+American+Chapter+of+the+Association+for+Computational+Linguistic) 12. Shaw, K. E., White, A. M., Moreton, E., & Monrose, F. 2014. Emergent faithfulness to morphological and semantic heads in lexical blends. In Proceedings of the annual meetings on phonology (Vol. 1, No. 1). [↗](https://doi.org/10.3765/amp.v1i1.45) 13. Thurner, Dick. 1993. Portmanteau dictionary: Blend words in the English language, including trademarks and brand names. Jefferson, NC: McFarland & Co. [find](https://scholar.google.com/scholar?q=Thurner%2C+Dick%2E+1993%2E+Portmanteau+dictionary%3A+Blend+words+in+the+English+language%2C+including+trademarks+and+brand+names%2E+Jefferson%2C+NC%3A+McFarland+%26+Co%2E) *Updated 1 November 2026*