📄 An Analysis of English Grammar in the Perspective of Nahwu

Summary

This qualitative study by Muh Shofiyudin, Hamidaturrohmah and co-authors contrasts English grammar with nahwu, the Arabic grammatical tradition. The motivation is pedagogical: an earlier questionnaire found that madrasa students in Kudus handle nahwu well but struggle with English grammar. The authors hope that mapping English onto categories these learners already know will reduce their errors.

The work is a library study with no corpus or computational component. Sources include Azar, Crystal, Frank and Bloomfield on the English side and the classical texts Jurumiyyah and Mutammimah on the Arabic side. The method has three steps: describe and classify each grammar, check that the categories are comparable, then compare them. English and Arabic experts validated the results.

The study covers two areas: word classes and verb tense. English has eight parts of speech, while Arabic has three kalimah classes. English tense is treated as a grid of time and aspect, while Arabic verb forms are organised by time only. The main value is a practical correspondence table for bilingual learners.

Key points

• Word classes: English has 8 parts of speech (noun, pronoun, verb, adverb, adjective, preposition, conjunction, interjection). Arabic kalimah has 3 classes: isim, fi'il and huruf.
• How the classes map:
  • Isim covers nouns, pronouns (personal and relative), adjectives and adverbs of time and place.
  • Fi'il corresponds only to verbs.
  • Huruf covers auxiliaries, prepositions and conjunctions.
• English tense: the paper counts 16 tenses, built as 4 time frames (present, past, future, past-future) crossed with 4 aspects (simple, continuous, perfect, perfect continuous).
• Arabic tense: verbs vary only by time (zaman: past, present, future), expressed with just two forms, fi'il madhi and fi'il mudhari'. Arabic has no separate concept of "tenses".
• Shared features:
  • Both languages have named word-class systems.
  • Both use different verb forms for different times.
  • Both distinguish verbal from nominal sentences.
  • Both rely on extra words to build negatives and questions (English auxiliaries, Arabic particles for negation and interrogation).

Interesting ideas

• I1. The paper does not treat the 16 English tenses as a flat list; it rebuilds them as the product of two independent axes (time × aspect), which makes the system compact and easy to check. — For Mova (tokens): Morphology could output time and aspect as separate features rather than a single tense label. Downstream stages could then reason about each axis on its own.
• I2. Comparing a 3-class system with an 8-class one produces an explicit many-to-one mapping table (for example, isim absorbs nouns, pronouns, adjectives and some adverbs), which documents exactly where the two categorisations disagree. — For Mova (tree): The same kind of table could be written for UD dialect conversion between treebanks with coarser or finer tagsets. It would show which conversions lose information and which can be undone.
• I3. The authors argue that Arabic classifies words by basic form alone, whereas English classes mix form and syntactic function; keeping these two criteria apart clarifies why category counts differ. — For Mova (tree): This supports keeping lexical category (UPOS) strictly separate from syntactic role (deprel). Words like adverbial nouns could then be tagged by form and labelled by function, without the two being blended.
• I4. Negation and question-forming particles in Arabic are grouped with English auxiliaries as one functional slot, defined by what they do rather than by their morphology. — For Mova (frames): Speech-act detection could key on a cross-linguistic "operator" slot covering auxiliaries, negators and question markers, rather than on specific lexical items.
• I5. The analysis adds a final step in which subject experts from both traditions validate the contrastive results using indicator-based sheets. — For Mova (learn): Silver-data checks could include a similar structured checklist that a human or LLM teacher fills in per category. This would make disagreements traceable to specific grammatical indicators.