Morpheme to Word Ratio Calculator
Compute morphemes per word — a measure of morphological complexity. Used by linguists and clinicians to compare child language development, second-language acquisition stages, and morphological typology across languages.
Last updated: May 2026
Compare with similar
About this calculator
Morpheme-to-word ratio = morphemes / words. A morpheme is the smallest meaningful unit in a language — "cats" has 2 morphemes (cat + plural -s); "rerun" has 2 (re- + run); "unhappiness" has 3 (un- + happy + -ness); "talked" has 2 (talk + past -ed). Variables: morphemes is the total count in your sample; words is the corresponding word count. Edge cases: words must be > 0; morpheme counting requires linguistic analysis — automated tools (CLAN, FreeMorph, Stanza) help but require manual review. Reference values vary by language type: analytical languages like Mandarin or Vietnamese hover near 1.0 morpheme/word (each word is typically a single morpheme); English averages ~1.3–1.5 in conversational speech (most words are monomorphemic but inflections push the average up); German and Russian ~1.8–2.5 (more inflectional complexity); agglutinative languages (Turkish, Finnish, Japanese, Korean) can reach 2.5–4 (single "words" combine many morphemes); polysynthetic languages like Inuktitut or Mohawk can routinely hit 5–10+ morphemes per "word" because what English would express in a sentence is packed into one polysynthetic word. The metric is central to language-typology classification: Joseph Greenberg's 1960 paper formalised the "Index of Synthesis" using this ratio. In child language development, Mean Length of Utterance in Morphemes (MLU-M) is a classic clinical metric — Roger Brown's 1973 stages defined developmental milestones from MLU-M = 1.0 (single-word stage, ~12-18 months) up through MLU-M = 4.0+ (complex sentence stage, ~age 4+). Lower MLU-M than age-typical can suggest language delay or specific language impairment.
How to use
Example 1 — Typical English sample. A short paragraph contains 28 morphemes across 20 words. Enter Number of Morphemes = 28, Number of Words = 20. Ratio = 28 / 20 = 1.4 morphemes per word. ✓ Typical of conversational English — most words are monomorphemic (the, of, that, is), with some complex words (running = 2 morphemes; happily = 2; unhappy = 2) pushing the average up. Same sample in a polysynthetic language might come out at 4–6 morphemes per "word". Example 2 — Child language sample (early speech). A 3-year-old produces 50 morphemes across 35 words in a play session. Enter 50, 35. Ratio (MLU-M) ≈ 1.43. ✓ Below average for age 3 (typical MLU-M ~3.0–3.5 by age 3), suggesting either an early-developing child or possible language delay; a speech-language pathologist would investigate further with structured assessments. Single-word stage children (12-18 months) have MLU-M near 1.0; by age 5 most children reach 4.0+. Brown's stages tie specific morphological structures to MLU-M thresholds (Stage I: 1.0-2.0; Stage V: 4.0+).
Frequently asked questions
What's the difference between a free morpheme and a bound morpheme?
Free morphemes can stand alone as words: "cat", "run", "happy", "the". Bound morphemes can't exist independently and must attach to a free morpheme: prefixes (re-, un-, dis-), suffixes (-s, -ed, -ing, -ness, -ly), inflections (verb endings, plural markers, case markers). English example: "unhappily" = un- (bound prefix) + happy (free root) + -ly (bound suffix) = 3 morphemes. Most English words are monomorphemic free roots ("dog", "house", "tree"), with the most common bound morphemes being plural -s, possessive 's, verb tense markers (-s, -ed, -ing), comparative -er and superlative -est, and a handful of derivational suffixes (-er, -ness, -ly, -tion, -ment). Different languages have very different ratios of free to bound morphemes; agglutinative languages have many bound morphemes stacked together (Turkish "evlerimden" = ev- "house" + -ler- plural + -im- "my" + -den ablative = "from my houses", 4 morphemes in one word).
How is morpheme ratio used in child language assessment?
Mean Length of Utterance in Morphemes (MLU-M) is one of the oldest and most-used clinical measures of language development. Roger Brown (1973) defined five developmental stages by MLU-M: Stage I (1.0–2.0, ~12–26 months) — single-word and early two-word utterances; Stage II (2.0–2.5, ~27–30 months) — emerging morphology like plural -s, present progressive -ing; Stage III (2.5–3.0, ~31–34 months) — more morphology, auxiliary verbs; Stage IV (3.0–3.75, ~35–40 months) — complex sentences with conjunctions; Stage V (3.75–4.5+, ~41+ months) — embedded clauses, advanced morphology. MLU-M correlates well with overall language ability through about age 4–5; after that, structural complexity (clause embedding, syntactic variety) becomes more informative than mere length. For clinical use, MLU-M is computed from a sample of ~50–100 spontaneous utterances; reliable scoring requires trained transcription and careful morpheme boundary decisions (some are contentious — is "going to" 2 morphemes or 1? Different scoring guidelines differ).
Why do languages differ so much in morpheme ratio?
Different morphological strategies. Analytic languages (Mandarin Chinese, Vietnamese, Yoruba) tend toward 1 morpheme per word — they express grammatical relationships through word order and function words rather than inflections. The Mandarin sentence "我 看 书" (wǒ kàn shū = I read book) has 3 words, each a single morpheme; tense is inferred from context or marked with separate aspect particles. Synthetic languages (English, German, Spanish, Russian) combine morphemes within words via inflections. Agglutinative languages (Turkish, Finnish, Japanese, Korean) stack many morphemes in transparent sequences within single "words"; Finnish "talossanikinkohan" = talo "house" + -ssa locative + -ni "my" + -kin "also" + -ko interrogative + -han particle = 6 morphemes in one word. Polysynthetic languages (Mohawk, Inuktitut, many Indigenous languages of the Americas and Siberia) pack what other languages express in full sentences into single polymorphemic words. The classification isn't absolute — languages have multiple strategies — but morpheme ratio is one quantitative way to compare across the spectrum.
What are the most common mistakes counting morphemes?
The first is treating phonological segments (sounds) as morphemes; "string" has 5 phonemes but is 1 morpheme. The second is missing zero morphemes — "sheep" (singular) and "sheep" (plural) have the same form but different morpheme counts in some analyses (1 vs 2 with a "zero plural" morpheme). The third is over-counting suppletive forms — "went" is sometimes analysed as go + past, but actually go and went are distinct lexical items, so most analyses count "went" as 1 morpheme. The fourth is mistakenly splitting bound roots; Greek-derived English words like "biology" (bio + logy) have 2 morphemes despite both being bound roots, while "telephone" is similarly bi-morphemic. The fifth is inconsistency in handling contractions; "don't" is sometimes 1 morpheme, sometimes 2 (do + not), and the choice affects MLU-M calculations for child language. The sixth is forgetting that morpheme analysis depends on the language's morphological system; what's 1 morpheme in English might be 3 in another language's analysis.
When should I not use this calculator?
Skip it for languages where morpheme boundaries are linguistically contested (Mandarin "compounds" like 电视 diànshì "television" = "electric+vision" — 1 or 2 morphemes?). Use language-specific conventions. Don't use it without consistent morpheme-counting rules; the same sample can yield different ratios depending on whether contractions, derivational vs inflectional morphemes, and zero morphemes are treated separately. It's the wrong tool for evaluating fluency or vocabulary breadth; those need different metrics (lexical diversity, type-token ratio, vocabulary size). Avoid it for assessing children with non-mainstream dialects without using dialect-appropriate norms; Standard American English MLU-M norms may not apply to AAE-speaking children, who follow their own developmental milestones. Don't use it for languages with non-Indo-European morphology without consulting language-specific developmental norms; child language milestones (and the morphemes that mark them) differ substantially across language families. Finally, for clinical diagnosis, MLU-M is one input — combine with comprehension testing, narrative analysis, and standardised assessments.