How To Calculate Mlu | A Core Linguistic Metric

Understanding how language develops in children is a fascinating and fundamental area of study. MLU provides a powerful, quantifiable metric for tracking a child’s linguistic growth, offering insights beyond simple word counts. This measurement helps educators and clinicians assess a child’s developing grammatical complexity and track their progress over time.

Understanding Mean Length of Utterance (MLU)

Mean Length of Utterance (MLU) serves as a fundamental measure in child language development research and clinical assessment. Roger Brown, a pioneering psycholinguist, introduced MLU in 1973 as a reliable indicator of grammatical complexity in children’s speech. Unlike merely counting words, MLU accounts for the internal structure of words, reflecting a child’s mastery of morphology and syntax.

MLU offers a more refined assessment of linguistic maturity than simply tallying the number of words a child uses. A longer MLU generally correlates with more grammatically complex speech, as children begin to incorporate inflections, derivations, and more sophisticated sentence structures. This metric is widely applied in developmental linguistics and speech-language pathology to monitor typical development and identify potential language delays.

Key Components: Utterances and Morphemes

Accurate MLU calculation depends on a precise understanding of its two core components: utterances and morphemes. These units form the basis of the entire measurement process.

Defining an Utterance

An utterance is a natural unit of speech, often delimited by pauses, a change in speaker, or a clear shift in communicative intent. It can be a single word, a phrase, or a complete sentence. The key is that it represents a single, coherent thought or communicative act.

Consistent segmentation of a language sample into individual utterances is vital for reliable MLU scores. Researchers and clinicians apply specific criteria to determine utterance boundaries, ensuring uniformity across analyses. For instance, a child saying “Big doggie” followed by a pause, then “Run fast,” would typically be segmented into two utterances.

Identifying Morphemes

A morpheme is defined as the smallest unit of meaning in a language. Morphemes can be either free or bound. Free morphemes stand alone as words (e.g., “cat,” “run,” “happy”), carrying independent meaning.

Bound morphemes, conversely, cannot stand alone and must attach to a free morpheme to convey meaning. These include prefixes (e.g., “un-” in “unhappy”), suffixes (e.g., “-s” for plural in “cats”), and verb inflections (e.g., “-ed” for past tense in “walked”). Counting morphemes involves dissecting each word into its meaningful components, adhering to standardized rules.

Standardized MLU Calculation Rules

The rules for counting morphemes are critical for consistent and comparable MLU results. While variations exist, many researchers and clinicians follow guidelines adapted from Roger Brown’s work and subsequent refinements, such as those by Miller and Chapman. These rules ensure that MLU reflects a child’s developing grammatical system accurately.

General Morpheme Counting Guidelines

Applying these guidelines systematically is essential for precise MLU calculation:

• Single Words: Most single words count as one morpheme (e.g., “book,” “table,” “jump”).
• Inflections: Regular plural markers (-s, -es), possessive markers (-‘s), third-person singular present tense (-s), regular past tense (-ed), and present progressive (-ing) each count as an additional morpheme when correctly used.
• Derivational Morphemes: Prefixes (e.g., “un-“) and suffixes (e.g., “-ly,” “-ness”) that change a word’s meaning or grammatical category typically count as separate morphemes.
• Compound Words: Generally, compound words that function as a single lexical item (e.g., “birthday,” “cupcake”) count as one morpheme. This reflects their conventional use as a single concept.
• Irregular Forms: Irregular past tense verbs (e.g., “went,” “ate,” “slept”) and irregular plurals (e.g., “feet,” “mice”) count as one morpheme, as the morphological change is not a simple addition.
• Reduplications: Reduplicated words or sounds (e.g., “bye-bye,” “choo-choo”) typically count as one morpheme.
• Auxiliary Verbs and Catenatives: Contracted forms (e.g., “I’m,” “he’s,” “they’re”) are often counted as two morphemes (pronoun + verb). Catenatives like “gonna,” “wanna,” “hafta” are usually counted as one morpheme if they function as a single unit, but some systems count them as two if the underlying morphemes are clearly present. For Brown’s stages, they are generally counted as one.
• Fillers and Non-Words: Non-meaningful vocalizations, fillers (e.g., “um,” “uh”), and false starts are not counted as morphemes.
• Repetitions and Reformulations: If a child repeats a word or phrase for emphasis or self-correction, only the first instance is typically counted, unless the repetition adds new meaning.

Table 1: Morpheme Counting Examples

Word/Phrase	Morpheme Breakdown	Morpheme Count
dog	dog	1
dogs	dog + -s	2
walked	walk + -ed	2
going	go + -ing	2
went	went	1
cupcake	cupcake	1
bye-bye	bye-bye	1
unhappy	un- + happy	2
they’re	they + are	2

The Step-by-Step Calculation Process

Calculating MLU involves a systematic process, from collecting speech samples to applying the final formula. Each step requires careful attention to detail to ensure accuracy.

Data Collection and Transcription

The process begins with collecting a representative language sample from the child. This typically involves recording spontaneous speech during naturalistic interactions, such as play or conversation with a familiar caregiver. A sample size of 50 to 100 complete, intelligible utterances is often targeted to provide a robust measure. Accurate transcription of this audio or video recording is paramount, capturing every word and relevant non-verbal cue.

Segmenting Utterances

After transcription, the next step involves segmenting the continuous speech into individual utterances. This requires careful listening to the recording and reviewing the transcript. Utterance boundaries are determined by factors such as a pause of two seconds or more, a clear change in intonation, a new speaker taking a turn, or a distinct shift in the topic or communicative intent. Consistency in applying these segmentation rules across the entire sample is essential.

Counting Morphemes per Utterance

Once utterances are segmented, each word within every utterance is analyzed for its morphemic content. The standardized morpheme counting rules are applied meticulously. For example, in the utterance “The dogs are running,” “The” is 1, “dogs” is 2 (dog + -s), “are” is 1, and “running” is 2 (run + -ing), totaling 6 morphemes for that utterance. This step requires a thorough understanding of morphology and careful application of the established guidelines.

The American Speech-Language-Hearing Association provides resources and guidelines for speech-language professionals on language sampling and analysis, including MLU calculation. These resources help ensure consistent and reliable practices.

Applying the MLU Formula

The final step is to apply the Mean Length of Utterance formula:
MLU = Total Number of Morphemes / Total Number of Utterances

To illustrate, if a child produced 50 utterances, and the sum of all morphemes across those 50 utterances was 125, the MLU would be 125 / 50 = 2.5. This numerical value represents the average morphological complexity of the child’s speech sample.

Interpreting MLU Scores

MLU scores are not just numbers; they serve as a valuable developmental index, offering insights into a child’s linguistic progression. Roger Brown’s seminal work established a strong correlation between MLU and stages of grammatical development, known as Brown’s Stages.

A child’s MLU typically increases with age, reflecting their growing ability to produce longer and more grammatically complex sentences. Comparing a child’s calculated MLU to age-referenced norms helps identify if their language development is proceeding as expected. A significantly lower MLU for a given age might suggest a language delay or disorder, prompting further assessment.

It is important to understand that MLU primarily measures morphological and syntactic complexity. It does not directly assess other vital aspects of language, such as vocabulary size, semantic understanding, or pragmatic skills (social use of language). Therefore, MLU is best used as one component within a broader language assessment battery, providing a specific window into grammatical development.

Table 2: Approximate MLU by Age (English-speaking children)

Age (Months)	Approximate MLU Range
18-24	1.0-2.0
24-30	2.0-2.5
30-36	2.5-3.0
36-42	3.0-3.75
42-54	3.75-4.5
54+	4.5+

Practical Considerations and Tools

While manual MLU calculation offers a deep understanding of the underlying principles, practical application often involves dedicated tools. Software programs, such as those available through the Child Language Data Exchange System (CHILDES) project, specifically the CLAN program, can automate the MLU calculation process from transcribed language samples. These tools enhance efficiency and reduce the potential for human error in large datasets.

Regardless of whether calculation is manual or automated, consistency in transcription and application of morpheme counting rules remains paramount. Training for transcribers and regular checks for inter-rater reliability are essential to ensure the validity and comparability of MLU scores. This careful approach ensures that MLU continues to be a reliable and informative metric in language assessment and research.

The National Institutes of Health supports research into child development, including language acquisition, providing foundational knowledge for these assessment methods.