Assimilation, sometimes called consonant harmony, happens when a child changes one sound in a word to match or resemble another sound in the same word. A child might say “gog” instead of “dog” because the final /g/ pulls the initial /d/ toward a velar place of production. Both velar and nasal assimilation typically resolve by age 3, so if the pattern persists beyond that point, it’s a reasonable target for intervention. Several evidence-based approaches can address it effectively.
What Assimilation Sounds Like
Assimilation is a phonological process, not an articulation error on a single sound. The child can often produce the target sound correctly in other words or contexts, but within certain words, one sound “pulls” a neighboring sound to match it. The two most common subtypes are velar assimilation and nasal assimilation.
In velar assimilation, a non-velar sound changes to a velar sound because of a nearby /k/ or /g/. “Duck” becomes “guck,” or “take” becomes “kake.” In nasal assimilation, a non-nasal sound shifts to a nasal because of a nearby /m/ or /n/. “Candy” might become “nandy.” The key diagnostic clue is inconsistency: the child can produce the replaced sound in words where no competing sound triggers the change.
Identifying the Pattern
Targeting assimilation starts with confirming that it’s actually happening as a pattern, not a coincidence across a few words. The most widely used standardized tool is the Goldman-Fristoe Test of Articulation, which surveys by far the majority of practicing SLPs report using routinely. About 30% of clinicians supplement it with the Khan-Lewis Phonological Analysis, which specifically codes phonological processes like assimilation from the same word sample.
Beyond standardized testing, a thorough phonological analysis of a single-word sample lets you calculate both the type and frequency of each error pattern. This matters because assimilation often co-occurs with other processes like fronting, stopping, or final consonant deletion. You need to know which patterns are most frequent and most impactful on intelligibility before deciding where to start. A connected speech sample adds context, since some children show assimilation more heavily in longer utterances where coarticulation demands increase.
Dynamic assessment can also help. Using a structured stimulability probe, you can test whether the child produces the target sound correctly when given models, cues, or simplified phonetic contexts. A child who is stimulable for the sound in isolation but loses it in words with competing sounds is showing a classic assimilation profile.
Minimal Pairs Therapy
Minimal pairs is one of the most direct ways to target assimilation because it highlights the communicative cost of the error. You select word pairs where the child’s assimilated production creates a homonym. If a child says “gog” for “dog,” you pair “dog” and “gog” (or a real-word near-equivalent) so the child experiences the breakdown firsthand.
In practice, the clinician places multiple pictures of both words on the table. The child names the target word, and if they produce the assimilated version, the clinician picks up the wrong picture. This creates a natural communicative consequence: the child said “gog,” so the clinician reached for the wrong card. Over repeated trials, the child learns to differentiate the sounds to get the intended result. The technique works because it makes the phonological contrast meaningful rather than abstract.
For velar assimilation specifically, you’d build pairs where the initial sound contrasts with the velar that’s triggering the error. Words like “tea” vs. “key,” “top” vs. “cop,” or “dome” vs. “gome” put the alveolar-velar contrast front and center. For nasal assimilation, pairs might contrast an oral consonant with the nasal version the child is substituting.
The Cycles Approach
The Cycles Approach is particularly well suited for children who show multiple phonological processes at once, which is common when assimilation is present. Rather than drilling one pattern to mastery before moving on, the clinician cycles through several target patterns over a set period, then loops back.
Each session follows a consistent structure: a brief review of the previous session’s target, auditory bombardment (where the child listens to a list of words containing the target pattern while the clinician reads them aloud, sometimes with slight amplification), production practice activities, stimulability probing for upcoming targets, and a phonological awareness component. The session closes with another round of auditory bombardment using the same word list.
For assimilation, you’d select target words where the child’s error pattern is active and place them in facilitative phonetic environments, meaning word shapes where the correct production is easiest. For a child with velar assimilation who turns initial /t/ into /k/ before a final velar, you might start with words where the target alveolar sound is surrounded by vowels that naturally encourage tongue-tip contact. Each pattern gets targeted for a set number of sessions (often two to three) before cycling to the next pattern, with the full cycle repeating until the pattern drops below a threshold.
Using Complexity to Drive Generalization
The complexity approach takes a counterintuitive angle: instead of starting with the easiest version of the error, you target harder linguistic structures first. The rationale is that mastering complex targets causes improvements to cascade down to simpler ones without direct treatment. A meta-analysis of treatment studies found that exposure to complex input not only improved production of those complex sounds (with a large effect size of 1.08) but produced even larger gains on untreated simpler sounds (effect size of 2.69).
Applied to assimilation, this might mean targeting words where the assimilation occurs in a consonant cluster rather than a simple onset, or choosing sounds the child has never produced correctly (two “new” sounds in a minimal pair) rather than sounds the child sometimes gets right. Research from multiple studies has shown that pairing two unacquired sounds leads to broader generalization than pairing a new sound with one the child already has.
This approach works best when you’re confident the child’s phonological system is ready for reorganization and when you want to maximize efficiency across multiple error patterns. It’s not the first choice for every child, but when assimilation co-occurs with several other active processes, targeting a complex form of the pattern can knock out simpler errors along the way.
Choosing Target Words Carefully
Regardless of which approach you use, word selection matters enormously when targeting assimilation. Because assimilation is triggered by the presence of a competing sound in the same word, you need to think carefully about which words to start with and which to build toward.
Early targets should place the correct sound in a facilitative position. For velar assimilation, this means choosing words where the alveolar target is in the initial position and the triggering velar is absent or far away. “Top,” “ten,” and “two” let the child practice the alveolar without velar interference. Once accuracy is solid, introduce words where both sounds coexist but the alveolar comes first and the velar is final: “tack,” “tick,” “take.” These are the words where assimilation actually occurs, so they represent the true therapeutic target.
For nasal assimilation, the same logic applies. Start with words where the oral consonant appears without a nearby nasal, then systematically introduce words that contain both the oral target and a nasal sound. The goal is to build the child’s ability to maintain two different manner categories within a single word.
Moving From Words to Conversation
A common concern is whether gains at the word level will transfer to spontaneous speech. Research on phonological process intervention suggests that for most children, system-level changes in single words do carry over to conversational speech without requiring a separate phase of treatment focused on spontaneous production. The phonological reorganization that happens during structured practice tends to ripple outward naturally.
That said, structured bridging activities can speed the process. Once a child is producing target words accurately in isolation, embed those words in carrier phrases (“I see a ___”), then short sentences, then story retells. Games that require repeated naming in a semi-spontaneous context, like “Go Fish” with target word cards, create a middle ground between drill and free conversation. Monitoring connected speech samples every few weeks lets you track whether the pattern is genuinely resolving or just suppressed during structured tasks.
When Assimilation Co-Occurs With Other Patterns
Assimilation rarely shows up alone. Studies of phonological error patterns in young children consistently find it alongside fronting, stopping, final consonant deletion, and cluster reduction. When deciding which pattern to target first, consider the frequency of each pattern, its impact on intelligibility, and whether the child is stimulable for the sounds involved.
Assimilation is sometimes a good early target because the child already has the sound in their inventory. They can produce /d/ and /g/ independently; they just can’t keep them separate in the same word. This makes it different from a pattern like stopping, where the child may genuinely lack the fricative sounds. Because assimilation is a sequencing problem more than a motor problem, many children resolve it relatively quickly once they become aware of the contrast.
Between ages 3 and 4, the overall use of phonological processes drops by roughly 50% in typically developing children. If a child is right at age 3 and assimilation is one of several active patterns, a cycles-based approach that rotates through all of them may be more efficient than isolating assimilation for intensive work. If the child is older and assimilation is one of the last remaining patterns, a focused minimal pairs or complexity approach can resolve it faster.