Semantic Equivalence Recognition: an undeclared operation in English language assessment
- 1Breakthrough English
Abstract
English language assessments require candidates to recognise that two differently worded expressions carry the same proposition. This paper adopts the name Semantic Equivalence Recognition for that operation and sets out evidence that it is operative in live items across boards and instrument types, and that it is not named as a receptive construct in the proficiency frameworks that govern English language instruction. The operation itself is not new. It was operationalised for first-language reading assessment in 1979, and it was named, given an item architecture, patented, tested diagnostically with a group of English as a second language learners, and published in a major educational technology journal between 2002 and 2006. That record is set out here in full, because the argument depends on it. The claim made is narrow: a construct with that pedigree is absent from the frameworks governing English language teaching, absent from the preparation literature for the examinations that depend on it, and has never been carried into high-stakes English language proficiency assessment. The paper also documents a point at which the question of whether the operation is taught was formally removed from an awarding body's own validation framework, and argues that disclosure obligations in assessment codes of practice are pitched at content coverage and item format and do not reach the level of the cognitive operation.
Related
- Bidirectional cross-scale detail-semantic collaborative network (BCDS-Net) for vision-based waste recognition
- Do Neuromodulator-Inspired State Prompts Generalize Across LLM Families? A Publicly Preregistered Audit of Output Length, Semantic Breadth, and Instruction Equivalence
- Hierarchical multimodal semantic fusion for few-shot reader action recognition in reading environments
- Contextual Semantic Pattern Recognition for Mental Health Monitoring Using ClinicalBERT Transformers.
- Semantic Context Alignment with Open-World Learning for Memory-Driven Human Activity Recognition
BibTeX
@misc{kocsis2026semantic,
title = {{Semantic Equivalence Recognition: an undeclared operation in English language assessment}},
author = {Balint Kocsis},
howpublished = {Zenodo},
year = {2026},
month = oct,
doi = {10.5281/zenodo.23450101},
url = {https://zenodo.org/records/23450101},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Semantic Equivalence Recognition: an undeclared operation in English language assessment
AU - Balint Kocsis
PY - 2026
DA - 2026/10/18
PB - Zenodo
DO - 10.5281/zenodo.23450101
UR - https://zenodo.org/records/23450101
AB - English language assessments require candidates to recognise that two differently worded expressions carry the same proposition. This paper adopts the name Semantic Equivalence Recognition for that operation and sets out evidence that it is operative in live items across boards and instrument types, and that it is not named as a receptive construct in the proficiency frameworks that govern English language instruction. The operation itself is not new. It was operationalised for first-language reading assessment in 1979, and it was named, given an item architecture, patented, tested diagnostically with a group of English as a second language learners, and published in a major educational technology journal between 2002 and 2006. That record is set out here in full, because the argument depends on it. The claim made is narrow: a construct with that pedigree is absent from the frameworks governing English language teaching, absent from the preparation literature for the examinations that depend on it, and has never been carried into high-stakes English language proficiency assessment. The paper also documents a point at which the question of whether the operation is taught was formally removed from an awarding body's own validation framework, and argues that disclosure obligations in assessment codes of practice are pitched at content coverage and item format and do not reach the level of the cognitive operation.
ER -