Language Development Disorders

Abstract

Language development disorders are a type of language disorder in which spoken language fails to develop at the expected rate in childhood, despite adequate hearing, non-verbal intelligence in the normal range, and no other condition that would explain the difficulty. The dominant form is developmental language disorder (DLD), long studied as specific language impairment (SLI). It is common, affecting roughly seven percent of children, persistent into adulthood, and marked most visibly by difficulty with grammar and with learning and retrieving words. This article traces how the category is defined and named, its grammatical signature and the nonword-repetition marker, its heritable and procedural-learning causes, how prevalence depends on the diagnostic threshold, and how the disorder is assessed. Three interactive demonstrations let a reader manipulate tense marking, repetition span, and the diagnostic criterion.

Keywords: developmental language disorder, specific language impairment, tense marking, nonword repetition, CATALISE

A language development disorder is a failure to acquire language typically, apparent from early childhood and out of keeping with a child's age, hearing, and non-verbal ability. It is not a disorder of the vocal apparatus or of the ear but of the language system itself, which is why a child with clear articulation and normal hearing can still struggle to combine words into grammatical sentences or to learn and retrieve vocabulary. The category is defined developmentally, in contrast with the acquired language disorders in which established language is lost after brain injury; both belong to the broader MeSH class of language disorders, but the developmental disorders arise as a system that never reaches typical function rather than one that was damaged (Leonard, 2014). Its cognitive interest is that language can fail to develop while the rest of cognition proceeds broadly normally, which implies that language acquisition draws on at least partly specialised machinery.

Types of Language Development Disorders

MeSH files language development disorders as a direct child of the broader descriptor Language Disorders and, in turn, gives them one narrower descriptor of their own. The classification is an indexing hierarchy for the medical literature, not a theory of how language is built, so a single child may satisfy more than one label and the subtype below is not mutually exclusive with the receptive, expressive, and mixed profiles the clinical literature also uses. Table 1 lists the direct MeSH child of the descriptor; it has no article of its own on this site yet and so is shown without a link.

Note. The subtype is listed as MeSH classifies it. Its narrower descriptor does not yet have its own article on this site, so the term is listed without a link. MeSH indexing is orthogonal to the clinical receptive-expressive distinction, so the two schemes cut the category along different lines.

From Specific Language Impairment to Developmental Language Disorder

For most of the twentieth century the developmental disorder now called DLD was known as specific language impairment. Two commitments were built into that name, and both proved untenable. The word specific implied that language was impaired in isolation, and the diagnostic practice that grew up around it required a discrepancy between language ability and non-verbal intelligence, on the theory that a child whose language merely matched a low IQ did not have a language disorder proper (Bishop, 2006). Neither survived scrutiny. Deficits in the condition routinely extend beyond language to processing speed, motor skill, and implicit learning, so the impairment is not narrowly specific; and the IQ-discrepancy requirement excluded many children with genuine, disabling language difficulties while predicting neither their prognosis nor their response to intervention (Leonard, 2014).

The current terminology was settled by CATALISE, a structured multinational Delphi study among specialists convened precisely because the profusion of overlapping labels was impeding research and services. Its first phase agreed criteria for identifying which children have clinically significant language problems (Bishop et al., 2016). Its second phase chose developmental language disorder as the umbrella term, explicitly dropped the non-verbal IQ cut-off, and reserved a separate label, language disorder associated with X, for language difficulties that accompany a known biomedical condition such as autism or hearing loss (Bishop et al., 2017). The change is not cosmetic: by removing the discrepancy criterion it widened the recognised population to include children with low-average non-verbal ability who need, and benefit from, the same support.

The Grammatical Signature

The most studied feature of English-speaking children with DLD is a specific, prolonged difficulty with grammatical morphology, and in particular with the markers of finiteness — the inflections and forms that signal tense and subject-verb agreement. Where typically developing children pass quickly through a stage of sometimes omitting these markers, children with DLD remain in it far longer, treating tense marking as optional in contexts where the adult grammar makes it obligatory. Mabel Rice, Kenneth Wexler, and Patricia Cleave named this the extended optional infinitive stage: the child produces bare, infinitive-like verb forms (he walk, she going) in place of finite ones (he walks, she is going), not at random but selectively, sparing morphemes that are not markers of finiteness (Rice, Wexler, & Cleave, 1995). Because the deficit is graded and morpheme-specific rather than an across-the-board failure, the proportion of obligatory contexts a child marks correctly, for a given morpheme at a given age, is one of the most sensitive clinical indicators of the disorder. The first demonstration lets a reader vary the morpheme and the child's age and compare the expected marking rate for a typically developing child against one with DLD.

Nonword Repetition as a Marker

If tense marking is the disorder's grammatical signature, nonword repetition is its most useful behavioural marker. The task is simple: the child hears an invented word of one to five syllables (ballop, woogalamic) and repeats it immediately. Performance falls as the item lengthens, because holding an unfamiliar sound sequence long enough to reproduce it draws on the phonological loop, the verbal component of working memory; children with DLD show a markedly steeper decline with length than their peers. Its value is twofold. First, because the stimuli are not real words, the task depends little on vocabulary knowledge or cultural exposure, so it is comparatively fair across dialects and languages. Second, it is strongly heritable: in a twin study, poor nonword repetition was substantially more concordant in identical than in fraternal twins and behaved as an inherited marker of language impairment, tracking the disorder within families even in children whose overt language had improved (Bishop, North, & Donlan, 1996). The second demonstration varies the syllable length and the group and plots the expected repetition accuracy.

Causes: Heritability and the Procedural Deficit Hypothesis

DLD runs in families and is highly heritable, but it is not a single-gene condition. Twin and family studies show substantial genetic influence on language impairment, yet the inheritance is polygenic and interacts with the environment, so no one gene or lesion accounts for the disorder and its expression varies widely between affected relatives (Bishop, North, & Donlan, 1996). This is the developmental counterpart of a point the acquired disorders make anatomically: language impairment need not correspond to damage in a single language module.

The most influential cognitive account of what is inherited is the procedural deficit hypothesis, which locates the core problem outside language proper. Michael Ullman and Elizabeth Pierpont proposed that DLD reflects an impairment of the brain's procedural memory system — the frontal-basal-ganglia circuitry that supports the implicit learning of rules and sequences — while declarative memory, which supports the learning of facts and arbitrary word-meaning pairings, is relatively spared (Ullman & Pierpont, 2005). On this view the grammatical difficulties follow naturally: the regularities of morphology and syntax are exactly the kind of sequential, rule-governed material that procedural memory normally acquires implicitly, and a child can partly compensate by routing rule learning through the intact declarative system, which is slower and more effortful. The hypothesis also predicts the non-linguistic difficulties in motor sequencing and implicit learning that helped discredit the older claim that the impairment was specific to language.

Identifying the Disorder and How Common It Is

Because DLD is defined behaviourally rather than by a biological test, deciding who has it means drawing a line on a continuous distribution of language ability, and where the line falls determines how many children are counted. The first rigorous epidemiological estimate came from a large kindergarten screening in the United States that applied an explicit diagnostic system and found specific language impairment in about 7.4 percent of five-year-olds (Tomblin et al., 1997; Tomblin, Records, & Zhang, 1996). Two decades later the population-based SCALES study in England, using a two-stage design and the broader, IQ-independent definition that CATALISE later endorsed, estimated that about 7.6 percent of children starting school have developmental language disorder of unknown origin, with roughly a further 2.3 percent showing language disorder associated with a known condition (Norbury et al., 2016). The crucial methodological finding was that prevalence swings substantially with the criterion, and especially with whether a non-verbal IQ cut-off is imposed: relaxing the old discrepancy requirement, as the field has now done, identifies more children as having a treatable disorder rather than a variant of low general ability. The third demonstration plots a cohort of children on non-verbal ability and language score and applies either the old discrepancy rule or the current IQ-independent rule, so the reader can see which children each criterion identifies.

Figure 1

The Uneven Profile of Developmental Language Disorder

Assessment

Because there is no biological marker, DLD is identified by structured testing of the language system, ideally combining norm-referenced instruments with markers that are robust to background and exposure. Standardised tests sample receptive language (comprehension) and expressive language (production) across levels — sounds, words, grammar, and connected discourse — and compare a child's scores with age expectations; a range of language tests exists for this purpose. Good practice supplements them with the two features that carry particular diagnostic weight: the marking of tense and agreement in obligatory contexts, and nonword repetition, both of which discriminate DLD from typical development better than vocabulary size and are less biased by exposure to the test language (Conti-Ramsden & Durkin, 2012). The central difficulty of assessment in the preschool years is separating a genuine, persistent disorder from the wide normal variation in the pace of early language and from the effects of limited exposure, which is why population screening uses two stages and why the diagnostic threshold is treated as a decision rather than a natural boundary (Norbury et al., 2016). Assessment also has to reckon with comorbidity: DLD frequently co-occurs with attentional, motor, and reading difficulties, so a language evaluation is rarely the whole clinical picture and a child identified with one condition warrants screening for the others (Bishop, 2006).

Worked Example

The dependence of a diagnosis on a threshold has a quantitative consequence worth working through, because even an excellent marker performs modestly as a population screen. Suppose nonword repetition is used to screen for DLD with a sensitivity of 0.90 (it correctly flags 90 percent of children who have the disorder) and a specificity of 0.80 (it correctly clears 80 percent of those who do not), against a population prevalence of 7 percent (Bishop, North, & Donlan, 1996). In a cohort of 1,000 children, 70 have DLD and 930 do not. The screen flags 70 x 0.90 = 63 of the affected children and misses 7; it also wrongly flags 930 x 0.20 = 186 unaffected children. So the total flagged is 63 + 186 = 249, of whom only 63 truly have the disorder — a positive predictive value of 63 / 249 = 0.25. Three-quarters of the children a good screen flags do not, on this arithmetic, have DLD.

The result is not a fault of the marker but of the base rate, and moving the base rate shows why targeted assessment works. Among children already referred to a clinic for language concerns, where the prevalence might be 25 percent rather than 7 percent, the same test yields 250 x 0.90 = 225 true positives against 750 x 0.20 = 150 false positives, a positive predictive value of 225 / 375 = 0.60 — more than double. This is the standard justification for a two-stage identification design: a cheap, exposure-fair screen first, then fuller assessment concentrated on the enriched group it identifies, which is precisely the structure the SCALES prevalence study adopted (Norbury et al., 2016). It also explains why the negative predictive value stays high throughout — a child the screen clears is very unlikely to have the disorder — so the screen is far better at ruling DLD out than at ruling it in.

Discussion

Developmental language disorder occupies an instructive place in cognitive science because it is the developmental complement to the aphasias. The acquired disorders show that adult language can be fractionated by damage; DLD shows that language can fail to assemble in the first place while the rest of cognition proceeds broadly normally, which is only possible if acquisition depends on mechanisms that are at least partly dedicated to it (Leonard, 2014). Yet the same evidence that motivates a degree of specialisation argues against a simple language module. The impairment reaches beyond language into implicit learning and sequencing, its inheritance is polygenic rather than Mendelian, and the procedural-deficit account explains the grammatical signature precisely by grounding it in a general learning system (Ullman & Pierpont, 2005). Language acquisition, on this reading, is specialised in what it builds but not in every mechanism it recruits.

The category also carries a lesson about classification that mirrors the acquired disorders. The older specific language impairment framework was tidy and teachable, and, in its strong form, wrong: the impairment was not specific, and the IQ-discrepancy rule it enforced excluded children who needed help without improving prediction (Bishop, 2006). The correction came not from a new intuition but from consensus methods and large, well-measured population samples, which redrew the boundary to match the evidence (Bishop et al., 2017; Norbury et al., 2016). The recurring shape of progress here is the same as elsewhere in the field: the phenomena are more graded, more heterogeneous, and more entangled with general cognition than the first clean models allowed, and the durable theory is the one that survives that complexity.

Current Directions

The consolidation of the diagnosis after CATALISE has shifted the research agenda from arguing about terminology toward closing the gap between what is known about DLD and how children are actually served. A prominent strand of current work is frankly critical: despite being one of the most common developmental conditions, DLD remains under-recognised by the public, under-identified in schools, and under-resourced relative to rarer disorders that attract more attention, and Karla McGregor has argued that the field fails affected children at each of these points and set out where recognition, identification, and intervention most need to improve (McGregor, 2020). A second strand pursues the procedural-deficit account with the tools of cognitive neuroscience and genetics, testing whether the implicated frontal-striatal learning circuitry and the polygenic risk it reflects can be tied more directly to the grammatical profile (Ullman & Pierpont, 2005). A third refines identification itself, seeking markers such as tense marking and nonword repetition that are fair across the many languages and dialects in which children are assessed, so that a diagnosis does not depend on exposure to a single test language (Conti-Ramsden & Durkin, 2012). The common thread is a move from defining the disorder toward reliably finding and helping the children who have it.

Common Misconceptions

Children with a language development disorder will grow out of it.

Some late talkers do catch up, but DLD as diagnosed is persistent: the difficulties typically continue into adolescence and adulthood, affecting literacy, schooling, and employment. The belief survives because early language varies so widely that a real disorder is easily mistaken for a passing delay, which is exactly why identification uses persistence and structured markers rather than a single early snapshot (Leonard, 2014).

A language development disorder means the child has low intelligence.

By definition DLD occurs in children whose non-verbal intelligence is in the normal range; the profile is uneven, with structural language impaired while reasoning is preserved. The current consensus deliberately dropped the requirement of a gap between language and IQ, because insisting on it excluded children with low-average ability who have genuine, treatable language difficulties (Bishop et al., 2017).

A single gene causes developmental language disorder.

DLD is strongly heritable but not Mendelian. Twin and family studies show substantial but polygenic genetic influence that interacts with the environment, so no one gene accounts for the disorder and its expression varies widely between affected relatives. The single-gene picture persists from well-publicised rare families, which are the exception rather than the rule (Bishop, North, & Donlan, 1996).

Glossary

CATALISE.

A multinational Delphi consensus study that agreed the criteria for identifying children with language problems and settled the term developmental language disorder.

Comorbidity.

The co-occurrence of another condition alongside a language disorder, such as attentional, motor, or reading difficulties.

Declarative memory.

The memory system for facts and arbitrary associations, including word meanings; relatively spared in DLD and available for compensation.

Developmental language disorder.

The current term for a persistent childhood difficulty acquiring language, not explained by hearing loss, low non-verbal IQ, autism, or a known medical condition.

Discrepancy criterion.

The abandoned diagnostic rule requiring a gap between language ability and non-verbal IQ before language impairment could be diagnosed.

Expressive language.

The production side of language: formulating and articulating words and sentences to convey meaning.

Extended optional infinitive.

A prolonged developmental stage in DLD in which tense and agreement marking is treated as optional in obligatory contexts, yielding bare verb forms.

Finiteness marking.

The grammatical marking of tense and subject-verb agreement, such as third-person -s and past-tense -ed, selectively difficult in DLD.

Heritability.

The proportion of variation in a trait attributable to genetic differences; substantial but polygenic for language impairment.

Nonword repetition.

A task requiring immediate repetition of an invented word; a heritable, exposure-fair behavioural marker of language impairment.

Obligatory context.

A sentence position where the adult grammar requires a particular morpheme, used to score whether a child supplies it.

Positive predictive value.

The proportion of children a screen flags who truly have the disorder; low when the base rate is low even for an accurate test.

Procedural memory.

The implicit memory system for rules, sequences, and skills; impaired in DLD on the procedural-deficit hypothesis.

Receptive language.

The comprehension side of language: recovering meaning from heard, read, or signed input.

Specific language impairment.

The former name for developmental language disorder, implying a language-specific deficit and a language-versus-IQ discrepancy now abandoned.

Key Researchers

Dorothy V. M. Bishop (b. 1952). Emeritus professor of developmental neuropsychology at the University of Oxford; she led the CATALISE consensus that established the diagnosis of developmental language disorder and demonstrated nonword repetition as an inherited marker. Wikipedia - Faculty Page - ORCID

Laurence B. Leonard (contemporary). Distinguished professor of speech, language, and hearing sciences at Purdue University; his monograph on children with specific language impairment defined the field and its cross-linguistic evidence base. Wikipedia - Faculty Page - Google Scholar

Karla K. McGregor (contemporary). Director of the Word Learning Lab at Boys Town National Research Hospital; she has documented how services, schools, and public awareness fail children with developmental language disorder. Faculty Page - ORCID - Google Scholar

Courtenay Frazier Norbury (contemporary). Professor of developmental language and communication disorders at University College London; she led the SCALES population study that produced a defensible, IQ-independent prevalence estimate for the disorder. Faculty Page - ORCID

Mabel L. Rice (contemporary). Distinguished professor of advanced studies at the University of Kansas; with Kenneth Wexler she developed the extended optional infinitive account of grammatical tense-marking deficits in DLD. Faculty Page - ORCID

J. Bruce Tomblin (contemporary). Distinguished professor emeritus at the University of Iowa; he directed the epidemiologic study that fixed the kindergarten prevalence estimate for specific language impairment and built its diagnostic system. Wikipedia - Faculty Page - ORCID

Michael T. Ullman (b. 1962). Professor of neuroscience at Georgetown University; with Elizabeth Pierpont he proposed the procedural deficit hypothesis that grounds DLD in a general implicit-learning system. Wikipedia - Faculty Page - Google Scholar

Frequently Asked Questions

What is a language development disorder?
It is a childhood difficulty acquiring spoken language at the expected rate, out of keeping with the child's age, hearing, and non-verbal ability. It is a disorder of the language system itself rather than of hearing or of the motor production of speech (Leonard, 2014).

What is the difference between specific language impairment and developmental language disorder?
They name the same broad condition at different times. Specific language impairment was the older term; the CATALISE consensus replaced it with developmental language disorder, dropping the requirement of a gap between language and non-verbal IQ (Bishop et al., 2017).

How common is developmental language disorder?
It affects roughly 7 to 8 percent of children starting school, making it far more common than autism. The exact figure depends on where the diagnostic threshold and any non-verbal ability cut-off are set (Norbury et al., 2016).

What are the main signs of the disorder?
The most visible signs are difficulty with grammar, especially marking tense and agreement, and difficulty learning and retrieving words. Poor repetition of invented nonsense words is a reliable behavioural marker (Rice, Wexler, & Cleave, 1995).

Is developmental language disorder inherited?
It is strongly heritable and runs in families, but the inheritance is polygenic rather than due to a single gene, and it interacts with the environment. Nonword repetition behaves as an inherited marker within families (Bishop, North, & Donlan, 1996).

Does a language development disorder mean low intelligence?
No. By definition it occurs in children whose non-verbal intelligence is in the normal range, with an uneven profile in which structural language is impaired while reasoning is preserved (Bishop et al., 2017).

Why is nonword repetition used to identify the disorder?
Because the words are invented, the task depends little on vocabulary or cultural exposure, making it comparatively fair across languages, and children with the disorder show a steep decline in accuracy as the words lengthen (Bishop, North, & Donlan, 1996).

Can developmental language disorder be treated?
It is managed with targeted speech and language intervention, and accurate assessment guides that support. The disorder is persistent, so intervention aims to improve function and outcomes rather than to cure (Conti-Ramsden & Durkin, 2012).

References