How Do You Calculate MLU? A Guide to Mean Length of Utterance

The Mean Length of Utterance (MLU) is a vital metric in linguistics and child language development, offering a quantitative measure of a child's linguistic complexity. It's often used by speech-language pathologists, researchers, and educators to assess language acquisition and identify potential developmental delays. Understanding how to accurately calculate MLU is crucial for anyone working with language samples.

What is Mean Length of Utterance (MLU)?

MLU refers to the average number of morphemes a speaker produces in an utterance. A morpheme is the smallest meaningful unit in a language. This includes not just words, but also grammatical markers like plural '-s', past tense '-ed', possessive '-s', and prefixes/suffixes. A higher MLU generally indicates more advanced linguistic abilities, as it suggests the speaker is combining more morphemes into complex sentence structures.

Why is MLU Important?

• Developmental Benchmark: MLU provides a reliable indicator of a child's grammatical development, correlating strongly with age during early language acquisition.
• Clinical Assessment: It's a standard tool for diagnosing language disorders and tracking progress in intervention programs.
• Research: Linguists use MLU to study language development across different populations and contexts.
• Educational Planning: Teachers can use it to tailor language instruction to a child's current developmental stage.

Step-by-Step Guide to Calculating MLU

Calculating MLU requires careful transcription and a systematic approach to morpheme counting. Here's how to do it:

Step 1: Collect a Language Sample

The first step is to gather a representative sample of the speaker's spontaneous language. For children, this usually involves recording a natural conversation or play interaction. A typical sample size for MLU calculation is 50 to 100 consecutive, intelligible utterances.

• Goal: Capture natural, spontaneous speech.
• Setting: Comfortable, familiar environment.
• Tools: Audio or video recorder.

Step 2: Transcribe the Sample

Transcribe the recorded sample verbatim. This means writing down every word, sound, and pause as accurately as possible. For MLU, focus on the speaker whose language you are analyzing.

Step 3: Segment into Utterances

An utterance is a natural unit of speech, typically marked by pauses, changes in intonation, or a new topic. It doesn't always correspond to a grammatical sentence. For example, "Big dog." "Run fast!" "My turn." are all separate utterances.

• Rules of thumb: A new utterance begins after a pause of 2 or more seconds, a change in speaker, or a clear shift in meaning.
• Exclusions: Exclude unintelligible utterances, imitations (unless spontaneously produced), and rote speech (e.g., counting, singing).

Step 4: Count Morphemes in Each Utterance

This is the most critical and often the most challenging step. You need to count every meaningful unit. Here are general guidelines, though specific protocols (like Brown's Stages or SALT) might have slight variations:

• Free Morphemes: Count each word as one morpheme (e.g., "cat," "run," "happy").
• Bound Morphemes: Count grammatical inflections and derivations as separate morphemes attached to a base word.
  • Plural '-s': "cats" (cat + s = 2 morphemes)
  • Possessive '-s': "mommy's" (mommy + s = 2 morphemes)
  • Third-person singular '-s': "runs" (run + s = 2 morphemes)
  • Past tense '-ed': "walked" (walk + ed = 2 morphemes)
  • Present progressive '-ing': "running" (run + ing = 2 morphemes)
  • Derivational suffixes: "unhappy" (un + happy = 2 morphemes)
• Compound Words: Typically counted as one morpheme if the child uses them as a single lexical unit (e.g., "playground," "birthday"). However, some protocols might count them as two. Be consistent.
• Contractions:
  • Count as one morpheme if they are part of a word the child is learning (e.g., "don't," "can't").
  • Count as two morphemes if the child demonstrates understanding of the two separate parts (e.g., "I'm" for "I am"). This often depends on developmental stage and transcription protocol. For basic MLU, often "I'm" = 1 morpheme.
• Reduplications: Count as one morpheme (e.g., "bye-bye," "choo-choo").
• Fillers: Usually not counted (e.g., "um," "uh," "like").

Step 5: Sum Total Morphemes

After counting morphemes for each utterance, add them all up to get the total number of morphemes in your entire sample.

Step 6: Sum Total Utterances

Count the total number of valid utterances in your sample.

Step 7: Calculate MLU

Divide the total number of morphemes by the total number of utterances:

MLU = Total Morphemes / Total Utterances

Example Calculation:

Let's say a child produced the following utterances:

1. "Doggie run." (Doggie + run = 2 morphemes)
2. "My ball." (My + ball = 2 morphemes)
3. "Kitty's sleeping." (Kitty + 's + sleep + ing = 4 morphemes)
4. "I want cookie." (I + want + cookie = 3 morphemes)
5. "He runs fast." (He + run + s + fast = 4 morphemes)

• Total Morphemes: 2 + 2 + 4 + 3 + 4 = 15 morphemes
• Total Utterances: 5 utterances
• MLU: 15 / 5 = 3.0

Interpreting MLU Scores

MLU scores typically increase with age, especially from ages 18 months to around 5 years. There are established norms (e.g., Brown's Stages of Language Development) that correlate MLU with chronological age. For instance, an MLU of 2.0-2.5 might correspond to a child around 24-30 months, while an MLU of 4.0-4.5 might be seen in children 42-48 months old.

A significantly lower MLU for a child's age might indicate a language delay or disorder, prompting further assessment.

Limitations and Considerations

While powerful, MLU is not the only measure of language development:

• Complexity vs. Length: A high MLU doesn't always equate to high grammatical complexity. A child might produce long, repetitive utterances without varied syntax.
• Language-Specific: Morpheme counting rules can vary slightly between languages.
• Sample Size: An insufficient or unrepresentative sample can lead to inaccurate MLU.
• Beyond Age 5: MLU tends to plateau after age 5, as children's sentences become grammatically stable, so other measures become more relevant for older children and adults.

Conclusion

Calculating MLU is a fundamental skill for anyone assessing language development. By carefully collecting, transcribing, and segmenting language samples, and then accurately counting morphemes, you can derive a valuable metric that sheds light on a speaker's linguistic capabilities. Remember to use MLU in conjunction with other qualitative and quantitative assessments for a comprehensive understanding of language proficiency.