Parent-child interaction gives children thousands of opportunities to hear language, respond, take conversational turns, learn words in context, and discover how sentences communicate increasingly complex ideas.
For children with developmental language disorder (DLD), those interactions remain important—but the relationship is frequently misunderstood.
DLD is not caused by parents failing to talk enough, read enough, or provide a sufficiently “language-rich” home. It is a neurodevelopmental communication disorder affecting the learning, understanding, and use of language. NIDCD estimates that DLD affects approximately 1 in 14 children in kindergarten and notes that it can affect speaking, listening, reading, and writing into adulthood.
At the same time, parents can be taught specific language-support strategies that increase children's opportunities to practice vocabulary, grammar, conversation, and other communication skills in everyday settings.
That distinction is central:
Parent-child interaction does not explain why a child has DLD.
But carefully structured parent-child interaction can be part of effective intervention.
A 2025 systematic review examining 67 studies of parental input in children with or suspected of having DLD found that interactive aspects of communication—particularly parent responsiveness and turn-taking—were positively associated with children's language outcomes. The sheer quantity of parental language, by contrast, was not consistently associated with better outcomes.
A separate meta-analysis of 76 parent-training studies involving 5,848 children found that teaching parents language-support strategies was associated with improvements in children's language and communication and with substantial increases in parents' use of those strategies.
The most useful clinical question is therefore not:
“Are the parents talking enough?”
It is:
“Which interactions help this child learn the particular language structures that are difficult, and how can those interactions be incorporated into intervention?”
First, What Happened to the Term Specific Language Impairment?
The term specific language impairment, or SLI, appears throughout older research and remains common in search queries, university coursework, and older clinical resources.
Current terminology has changed.
In 2017, the multinational CATALISE consensus project recommended the term Language Disorder for language difficulties that create significant functional impairment and are associated with poor prognosis.
The term Developmental Language Disorder (DLD) was recommended when that language disorder is not associated with a known biomedical condition such as a genetic syndrome, acquired brain injury, or sensorineural hearing loss.
Importantly, DLD does not require:
an IQ-language discrepancy;
normal performance in every nonlanguage domain;
or the absence of co-occurring developmental difficulties.
CATALISE specifically concluded that DLD can coexist with conditions such as ADHD and that the presence of biological or environmental risk factors does not automatically exclude the diagnosis.
NIDCD has adopted the term developmental language disorder while acknowledging that DLD was historically called specific language impairment, developmental dysphasia, or language delay.
For older intervention studies, this article will retain the term SLI when that was the diagnosis used by the original researchers.
Otherwise, DLD is the preferred term.
What Is Developmental Language Disorder?
DLD affects a child's ability to learn and use spoken language despite adequate opportunities to learn language.
The difficulty can involve several areas.
A younger child may have difficulty learning words, combining words, using grammatical markers, or understanding instructions.
An older child may speak in sentences but struggle with complex syntax, word retrieval, figurative language, narratives, classroom instructions, reading comprehension, or written language.
NIDCD describes DLD as a neurodevelopmental condition resulting from complex interactions among genes, development, and environment. It often runs in families: NIDCD reports that approximately 50%–70% of children with DLD have at least one family member with a history of language difficulties.
DLD is therefore fundamentally different from the idea that a child simply did not receive enough verbal stimulation.
A highly responsive parent can have a child with DLD.
A child can grow up surrounded by books and conversation and still have DLD.
Intervention focuses on helping the child learn more efficiently despite the underlying language-learning difficulty.
Why Parent-Child Interaction Still Deserves Attention
If parents do not cause DLD, why is parent interaction studied so extensively?
Because children spend far more time communicating outside therapy than inside it.
A child receiving two 30-minute language sessions each week has many more communication opportunities during breakfast, play, dressing, travel, homework, bath time, books, family meals, and weekend activities.
Parent-mediated intervention attempts to use those ordinary contexts deliberately.
Instead of expecting a parent to become a substitute speech-language pathologist, the clinician teaches a limited set of strategies that can be used during interactions that would already occur.
For example, a parent may learn to:
follow the child's focus of attention;
wait for the child to initiate;
respond contingently;
expand an utterance;
provide repeated examples of a grammatical form;
create opportunities for communication;
or recast an incorrect sentence while preserving its meaning.
The objective is not to maximize the number of adult words.
It is to improve the quality and instructional usefulness of interaction.
What Does Current Research Say About Parent Input in DLD?
The 2025 systematic review by van Witteloostuijn and colleagues provides one of the clearest current summaries.
The researchers reviewed 67 studies examining parental language input in children ages 0–6 with or suspected of having DLD. Parental input was divided into three broad dimensions:
interactive features, such as responsiveness and turn-taking;
linguistic features, such as grammatical complexity and vocabulary;
and conceptual features, relating to the type of ideas communicated.
Interactive aspects were studied most frequently.
Across studies examining the relationship between parent input and children's language outcomes, responsive interaction showed the most consistent positive associations. Simply measuring how much parents talked did not show the same consistent relationship.
That result helps move the field away from simplistic prescriptions such as:
“Just talk to your child more.”
The more useful clinical emphasis is:
notice what the child communicates → respond to it → provide a developmentally appropriate model → give the child another opportunity to participate.
Association Is Not the Same as Cause
This distinction is essential.
If responsive parent behavior is associated with stronger child language performance, several explanations are possible.
Responsive interaction may support language learning.
Children with stronger language may also make it easier for parents to respond conversationally.
Parents may adjust how they speak because of the child's communication abilities.
Or several influences may operate simultaneously.
The 2025 review specifically reflects a body of observational as well as intervention research and should not be interpreted as proof that differences in parenting caused children's language outcomes.
This is particularly important in DLD because children can influence the interactions around them.
A child who rarely speaks gives the parent fewer opportunities to expand utterances.
A child who responds slowly may unintentionally invite adults to ask additional questions before they have time to answer.
A child who has difficulty understanding complex sentences may lead adults to simplify their language.
Parent and child behavior continually influence each other.
Parent-child interaction is therefore best understood as a dynamic system, not as a one-way flow of language from parent to child.
Where Does Vygotsky's Social Interactionist Theory Fit?
Lev Vygotsky's social interactionist theory offers one theoretical framework for considering supported language learning.
It can remain, but its role should be described accurately.
Vygotsky argued that learning and cognitive development occur within social interaction and introduced the concept now widely described as the zone of proximal development: the difference between what a learner can accomplish independently and what they can accomplish with appropriate support.
The related concept of scaffolding is commonly used in education and language intervention to describe temporary support that helps a child perform beyond their current independent level.
For example, a child may be unable to independently retell a story.
With a visual organizer and questions such as:
“Who was the story about?”
“What went wrong?”
“What happened next?”
the child may produce an organized narrative.
As performance improves, those supports can be reduced.
That framework is useful for thinking about intervention.
But Vygotsky's theory is not evidence that social interaction causes or cures DLD.
Modern DLD research comes from developmental science, genetics, psycholinguistics, intervention trials, and language-learning research.
Social interactionist theory can help explain how support may be structured during learning.
It should not replace current clinical evidence.
Parent-Child Interaction Is Not a Single Treatment
“Improve parent-child interaction” is too broad to serve as a clinical intervention.
Different treatment programs manipulate interaction in different ways.
A clinician may target:
responsiveness;
turn-taking;
language models;
recasts;
focused stimulation;
communication temptations;
questions;
waiting;
expansions;
grammar;
vocabulary;
or communication independence.
The effectiveness of any parent-mediated intervention depends partly on whether the strategy corresponds with the child's actual communication target.
A toddler who rarely communicates may require opportunities for intentional communication.
A preschooler with DLD who speaks frequently but omits grammatical morphemes needs a different strategy.
A school-age child struggling with complex sentences needs another approach again.
How Do SLPs Assess Parent-Child Interaction?
Parent-child interaction should not be evaluated through intuition alone.
ASHA describes several assessment methods that can provide functional information about language and the contexts in which communication occurs.
Naturalistic Observation
The clinician observes the child communicating with caregivers during play, conversation, routines, or another familiar activity.
The clinician may examine:
who initiates interactions;
how often the child responds;
how long conversational exchanges continue;
how the parent responds to the child's communication;
what happens when communication breaks down;
and whether the child's language changes under different levels of support.
ASHA describes systematic observation and contextual analysis as useful complements to standardized assessment because they show communication under different tasks and settings.
Language Sampling Provides More Detail Than a Global Interaction Rating
A naturalistic language sample can be particularly informative.
The clinician might record 10–20 minutes of parent-child play and analyze:
sentence length;
grammatical morphology;
verb use;
number of different words;
complex syntax;
questions;
narrative language;
communication functions;
or another measure relevant to the referral.
ASHA lists language sampling across free play, conversation, narrative, and expository contexts among valid assessment techniques and notes that measures such as mean length of utterance, lexical diversity, developmental sentence scoring, and clausal complexity can complement standardized testing.
This is more useful than simply concluding:
“Parent and child have good interaction.”
The clinician needs to know what the child actually does with language and what happens when the adult changes their response.
Parent Interaction Should Not Replace Comprehensive DLD Assessment
Observation of parent-child interaction is one source of information.
It cannot establish DLD by itself.
A comprehensive assessment may include:
case history;
hearing status;
standardized measures when appropriate;
language samples;
narrative assessment;
functional observation;
caregiver and teacher report;
dynamic assessment;
and analysis of communication across settings.
ASHA emphasizes the use of multiple converging sources of evidence, particularly because a single standardized score cannot describe the complete language profile.
For multilingual children, the limitations of relying on one standardized English assessment become even greater.
DLD Assessment in Multilingual Children Requires Additional Care
Multilingualism does not cause DLD.
NIDCD explicitly states that learning more than one language does not cause the disorder and is not harmful to a child who has DLD. A multilingual child with DLD will demonstrate an underlying language-learning difficulty across the languages they use, although the surface form of that difficulty will differ across languages.
ASHA recommends considering all languages and dialects used by the child and obtaining speech and language samples across languages when possible.
A standardized test normed on monolingual English speakers should not be translated and then interpreted as though the standard score remains valid.
ASHA specifically states that standard scores should not be reported when a test has been translated or administered to a population for which it was not intended.
Naturalistic parent-child interaction can be especially useful here because it gives clinicians an opportunity to observe communication in a language the family actually uses.
Dynamic Assessment Can Complement Parent-Child Observation
Dynamic assessment asks not only what a child knows now but how the child responds to teaching.
A common structure is:
test → teach → retest
The clinician may briefly teach a narrative structure, grammatical form, or novel language pattern and then examine how much support the child requires to learn and apply it.
ASHA describes dynamic assessment as a useful component of language assessment, particularly when clinicians need information about learning potential rather than performance relative to monolingual norms.
A 2022 systematic review found that dynamic assessment showed promise for distinguishing language disorder from language difference in multilingual children, although the authors noted small samples and methodological limitations in much of the evidence.
More recent evidence has strengthened the field.
A 2024 study involving 634 diverse first- through fifth-grade students reported greater than 90% sensitivity and specificity for a brief dynamic narrative assessment, with little evidence that scores differed systematically according to multilingual status or several other demographic variables.
A small 2026 feasibility study also reported promising classification results for a dynamic statistical-learning task in Spanish-English preschoolers, though the sample included only 18 children and requires replication.
Enhanced Milieu Teaching: What It Actually Is
Enhanced Milieu Teaching (EMT) is one evidence-supported naturalistic language intervention.
EMT is a naturalistic language intervention designed primarily for young children with emerging language.
It combines responsive adult interaction with intentional teaching procedures embedded within ordinary activities.
Typical EMT components include:
following the child's lead;
responding contingently;
expanding what the child says;
modeling language at an appropriate level;
using environmental arrangements to create communication opportunities;
and prompting communication when appropriate.
The important point is that EMT is not simply “creating a language-rich environment.”
It is a structured intervention with specific procedures.
ASHA includes milieu teaching among naturalistic or hybrid approaches used to support spoken-language development.
How Strong Is the Evidence for Enhanced Milieu Teaching?
The evidence base is substantial, although the populations studied are heterogeneous.
A recent systematic review and meta-analysis examined 46 EMT studies involving 1,590 children.
Thirty-nine studies met What Works Clearinghouse standards without reservations. Across the evidence base, EMT produced positive effects on child communication and language outcomes. The review also found larger effects when EMT was implemented by caregivers than when implementation was limited to therapists.
This is important because EMT has been studied across children with different developmental profiles—not only children with DLD.
For example, a randomized study by Kaiser and Roberts included 77 preschool children with intellectual disabilities, including children with Down syndrome and autism. Parents who were taught EMT used more intervention strategies at home, and several observational language outcomes favored the parent-plus-therapist condition over time.
Those findings support the parent-mediated model.
They should not be described as a trial specifically of DLD.
Newer EMT Research Is Moving Closer to DLD-Specific Populations
More recent studies are directly addressing children with DLD or at elevated risk for it.
A 2024 small randomized trial of the Parents Plus program trained parents of preschool children with DLD to use focused stimulation. Results provided preliminary evidence of improvement in vocabulary and morphosyntactic skills, while parents rated the intervention as acceptable. The sample included only 31 parent-child dyads, so larger trials are still needed.
An important 2026 randomized controlled trial is now examining an adapted intervention called Enhanced Milieu Teaching–Sentence Focused (EMT-SF).
The study includes 108 children at risk for DLD, beginning at approximately 30–31 months of age. Caregivers receive intervention coaching through telehealth, and the larger project is designed to examine progression from vocabulary into grammatical development.
This newer work helps strengthen the DLD-specific evidence base for caregiver-mediated naturalistic intervention.
What Might EMT Look Like During Everyday Play?
Suppose a child is playing with cars and currently produces mostly single words.
Child:
“Car.”
Instead of asking:
“What color is the car?”
the parent might respond:
“Red car.”
The child pushes it.
Parent:
“Car goes!”
The child reaches for another one.
The parent pauses briefly, giving the child an opportunity to communicate.
Child:
“More car.”
Parent:
“More cars. You want more cars.”
The interaction is responsive, but the adult is also providing repeated language models slightly beyond the child's current level.
For another child, the target may be grammatical:
Child:
“Dog running.”
Parent:
“The dog is running.”
The activity looks like ordinary play.
The clinician's target determines which language models are emphasized.
Environmental Arrangement Should Create Opportunity, Not Frustration
Your original EMT example suggested withholding a puzzle piece so the child has to request it.
That technique can create a communication opportunity, but it should be used carefully.
The purpose is not to frustrate a child until they speak.
A clinician may arrange materials so that communication becomes useful—for example, providing part of a preferred toy set and making the remaining materials visible and accessible through an appropriate request.
But communication can occur through:
speech;
gesture;
sign;
AAC;
pointing;
or another established modality.
The adult should recognize valid communication rather than require a particular spoken phrase before granting access.
For children with more developed language, environmental arrangement may be unnecessary altogether.
Their treatment may depend more heavily on carefully selected models, recasts, narratives, or explicit language instruction.
Conversational Recasts Have a Strong Evidence Base for Grammar
Conversational Recast Training is another evidence-supported intervention approach.
Recasts are one of the better-supported intervention tools for grammatical development.
A child produces an utterance containing an error:
“She run fast.”
The adult responds naturally:
“Yes, she runs fast.”
The adult has not asked the child to repeat the sentence.
Instead, the child's original meaning is preserved while the grammatical form is corrected.
A 2015 systematic review and meta-analysis by Cleave and colleagues examined 35 studies in which recasts played a central intervention role and concluded that the overall evidence supported recasting for improving grammatical development in children with language impairments.
Recasting is therefore more than a general suggestion to “model good grammar.”
In well-designed intervention, the clinician selects a specific grammatical target and creates many opportunities for that target to occur.
Recasts Work Best When the Target Is Defined
Suppose a child omits third-person singular -s:
“He walk.”
“She eat.”
“Daddy drive.”
The intervention target may be:
verb + -s when the subject is third-person singular
During treatment:
Child:
“The boy jump.”
Adult:
“Yes, the boy jumps.”
Child:
“Mommy cook.”
Adult:
“Mommy cooks.”
The recasts repeatedly highlight the same grammatical relationship.
This is different from correcting every grammatical error the child makes.
A child with DLD may have difficulty with several grammatical structures at once.
Effective treatment usually targets selected forms systematically.
Enhanced Conversational Recast Treatment Shows Direct Evidence in DLD
Several intervention studies using the older diagnostic term SLI provide direct evidence for recast treatment.
Meyers-Denman and Plante studied 16 children ages 4–5 with SLI receiving Enhanced Conversational Recast treatment targeting grammatical morphology.
Children improved significantly on their treated grammatical morpheme while an untreated comparison morpheme remained unchanged, supporting a treatment rather than simple maturation explanation. The study also compared one 30-minute treatment period with three 10-minute periods during the same day and found no clear difference between the schedules.
More recent research continues to refine dosage.
A 2025 scoping review of 32 publications concluded that effective recast interventions have used several different dosage configurations and that the optimal dosage remains unknown because intervention studies describe and manipulate dosage inconsistently.
Clinicians should therefore avoid claiming that one fixed number of recasts or one session schedule is universally superior.
The Child's Starting Point Can Affect Grammatical Treatment
A newly published 2026 study provides an interesting refinement.
Thirty-four preschool children with DLD received Enhanced Conversational Recast treatment.
Some children began treatment on a grammatical form they already used occasionally before moving to a form they rarely used. Others began immediately with the more difficult form.
Children who first received treatment on the partially acquired form subsequently made faster initial gains on the more difficult target.
The finding suggests that treatment sequencing may influence how efficiently children learn.
This reinforces a broader principle:
language intervention should begin from the child's actual linguistic system rather than a generic developmental checklist.
Focused Stimulation Is Another Parent-Friendly Strategy
Focused stimulation involves providing a high density of examples of one language target within an activity without demanding that the child imitate every model.
Suppose the target is the location word under.
During play:
“The car is under the chair.”
“Let's put the bear under the table.”
“I found your block under the box.”
“Uh-oh, the ball rolled under the couch.”
The target occurs repeatedly, but the interaction remains communicative.
ASHA includes focused stimulation among hybrid spoken-language intervention approaches.
The 2024 Parents Plus trial also used caregiver training in focused stimulation with preschool children with DLD and reported promising preliminary language outcomes.
.png)
.png)