How do you calculate MLU?

Count the morphemes in each intelligible utterance of a language sample, add them up, and divide by the number of utterances. MLU in words uses the same steps but counts words instead of morphemes.

How many utterances do you need for MLU?

Most clinicians use at least 50 consecutive intelligible utterances. Brown's original procedure used 100. Smaller samples give a less stable estimate.

What is a normal MLU for a 3-year-old?

Using the Miller and Chapman (1981) prediction equation, MLU = -0.548 + 0.103 x age in months, the expected MLU at 36 months is about 3.16 morphemes. Individual children vary widely around this value.

MLU Calculator

Paste a language sample, one utterance per line, and get the mean length of utterance in morphemes and in words, the matching Brown's stage, and the expected MLU for the child's age. Nothing leaves your browser.

MLU in morphemes (MLUm)

MLU in words (MLUw)

Utterances counted

Total morphemes

Total words

Brown's stage

Utterance-by-utterance counts

#	Utterance as counted	Words	Morphemes

What MLU measures

Mean length of utterance is the average number of morphemes (or words) a child uses per utterance in a spontaneous language sample. Roger Brown introduced it in 1973 as a simple index of grammatical development in young children: as children add grammatical endings, function words and longer phrases, their MLU rises. Speech-language pathologists use it in evaluations and progress notes for children roughly 18 months to 5 years old.

MLU in morphemes counts every meaningful unit, so dogs is two (dog + plural -s) and jumped is two (jump + past -ed). MLU in words simply counts words. Report which one you used; they are not interchangeable.

How to count morphemes (Brown's rules, simplified)

Count as separate morphemes	Count as one morpheme
Plural -s (cat/s)	Irregular past tense (went, saw, got, did)
Possessive 's (mommy's)	Compound words and proper names (birthday, choo-choo, Big Bird)
Third person -s (he run/s)	Diminutives (doggie, mommy)
Regular past -ed (jump/ed)	Catenatives (gonna, wanna, hafta)
Present progressive -ing (play/ing)	Ritual repetitions (no no no said as one unit)
Contractions (it's, I'm, we'll) — the calculator adds one automatically for the apostrophe part	Irregular plurals (men, feet, mice)

Also leave out: false starts and revisions (put them in parentheses, as in (the the) ball), fillers such as um, and utterances that are partly unintelligible. Type unintelligible words as x, xx or xxx and the calculator skips that utterance.

Using the calculator

Paste or type the child's utterances, one per line. If your transcript has speaker codes (C for child, E for examiner at the start of each line, as in SALT), only the child's lines are counted.
Mark bound morphemes with a slash: dog/s, walk/ed, eat/ing. The Mark -ing / -ed / -s for me button does a first pass automatically. It cannot tell bus from bug/s in every case, so read through the result and fix anything it missed or over-marked.
Apostrophe forms (mommy's, it's, don't) are counted automatically; you do not need a slash. Tick the box if your protocol counts don't, can't, won't as single morphemes, which many clinicians do when the child never uses do, can or will with a separate not.
Enter the child's age in months to compare with the expected MLU, then copy the summary into your report.

Brown's stages and MLU

Stage	MLU (morphemes)	Typical age	What appears
I	1.0–2.0	12–26 months	Single words, then two-word combinations (agent + action, action + object)
II	2.0–2.5	27–30 months	-ing, plural -s, in and on
III	2.5–3.0	31–34 months	Irregular past (went, fell), possessive 's, uncontractible copula, early questions and negatives
IV	3.0–3.75	35–40 months	Articles a and the, regular past -ed, embedded clauses begin
V	3.75–4.5	41–46 months	Third person -s, irregular third person, uncontractible auxiliary, contractible copula and auxiliary, compound sentences
Post-V	4.5+	47+ months	Complex sentences continue to develop

The age ranges are those usually cited from Miller (1981). They describe typical development, not cut-offs.

Expected MLU by age

Miller and Chapman (1981) fitted a straight line to MLU in morphemes for children aged 18 to 60 months: expected MLU = −0.548 + 0.103 × age in months. The calculator uses that equation when you enter an age. A few reference points:

Age	Expected MLU
24 months	1.92
30 months	2.54
36 months	3.16
42 months	3.78
48 months	4.40
60 months	5.63

Children of the same age differ a lot, and MLU also depends on the sampling context (play, narrative, conversation). Use the expected value as one piece of evidence, together with standardized tests and clinical judgment.

Frequently asked questions

What is the formula for MLU?

MLU = total number of morphemes ÷ total number of utterances. For MLU in words, divide total words by total utterances.

How many utterances should the sample have?

At least 50 consecutive intelligible utterances is the common clinical standard; Brown used 100. The calculator warns you when fewer than 50 are counted.

Is this the same as SALT's MLU?

It follows the same basic conventions (one utterance per line, slashes for bound morphemes, parentheses for mazes, x for unintelligible words), so a SALT-style transcript gives very similar numbers. SALT has many more codes and its own databases for comparison; this page is a quick check, not a replacement.

Is MLU useful for school-age children?

MLU rises quickly up to about age 5 and changes slowly after that, so it is less informative for older children. Clinicians often switch to measures such as MLU in C-units or T-units, or clausal density, for school-age language samples.

Is my sample stored?

No. The text you paste is processed by JavaScript in your browser and never sent anywhere.