For speech-language pathologists
Articulation word lists vs the words kids say most: what the data shows
Published · Updated
The short answer
In our own picture-based word lists, the problem wasn’t rare words but missing small ones. We checked 372 words from our articulation word lists and flash cards for S, R, L, K and TH against Wordbank’s parent-report data for 2½-year-olds and Dolch’s list of service words. Where the pictured words are on the 680-word toddler checklist, they’re well known: for S, R, L and K, a median of 86% to 91% of 2½-year-olds say them. But the S, R, L and K lists include just 11 of the 134 Dolch service words with those sounds, and none of the 25 with K. “Yes”, the most-said S word on the checklist (99%), isn’t on either S list, and “please” (98%) and “thank you” (98%) are missing from the L and K lists. TH was the exception, with 11 of 16. The case for adding these words is what children need every day, not a promise of faster progress.
Why we checked our own lists
A speech-language pathologist (SLP) we spoke with said articulation materials often use low-frequency words and ignore the order children develop in. Her example: there’s little point practicing “seal” if a child can’t say “so” or “said”. This page checks the first point: which words are on a list.
Our older free word lists were built around pictures. Their method note says the words are “early, concrete words that most children aged 4 to 8 can picture”, and every flash card needed a picture that clearly shows the word. Small words like yes, so and said can’t be drawn. So we counted what that format left out, using our own articulation word lists and flash card decks. We haven’t analyzed anyone else’s materials, and these numbers say nothing about them.
Finding 1: the small words are missing
Dolch’s service words are the non-noun words, such as said, so and soon, that Dolch found children meet most in books (Dolch, 1936). We used them as a marker of the small words children use all day. For each sound, we counted the Dolch words that contain it, and how many are on our picture lists and on our newer high-frequency lists.
| Sound | Dolch words with the sound | On our picture lists | On our high-frequency list | Examples missing from the picture lists (share of 2½-year-olds who say it) |
|---|---|---|---|---|
| S | 28 | 3 (saw, sleep, stop) | 25 | yes 99%, sit 92%, see 88%, some 78%, so 38%, said (not on the CDI) |
| R | 47 | 4 (around, four, red, run) | 37 | drink 94%, read 92%, here 83%, where 82%, for 63%, are 57% |
| L | 34 | 4 (blue, fly, light, yellow) | 28 | please 98%, play 95%, help 88%, look 87%, like 86%, all 74% |
| K | 25 | 0 | 24 | cold 95%, walk 91%, look 87%, like 86%, can 72%, come (not on the CDI) |
| TH | 16 | 11 | 16 | with 69%, then 34%, their 29%; three and together (not on the CDI) |
| S, R, L and K | 134 | 11 | 114 |
The 11 are generous to the picture lists: “saw” is pictured as the tool and “fly” as the insect. TH is the exception because our older TH list added voiced words such as the, this, they and there on purpose. It leaves out “three” on purpose too, because THR is a blend.
The same gap shows up among the most-said words on the toddler checklist. “Yes” is the most-said S word, “please” is tied for second among L words, and “thank you” is second among K words and first among TH words. Yet “yes” isn’t on our S picture lists, “please” isn’t on our L lists and “thank you” isn’t on our K lists.
Finding 2: the pictured words are familiar, where there’s data
We expected the picture lists to be full of rare words. Mostly they weren’t. For each sound, we looked up every picture-list word on the MacArthur-Bates Communicative Development Inventory (CDI): Words and Sentences, a 680-word checklist parents fill in, using Wordbank’s American English data for 30-month-olds (Frank et al., 2017; Frank et al., 2021).
| Sound | Picture-list words | On the CDI | Median: picture-list words | Median: all CDI words with the sound | Median: our high-frequency list | Of the 10 most-said CDI words with the sound, on the picture lists |
|---|---|---|---|---|---|---|
| S | 76 | 36 | 88% | 80% (137 words) | 84% | 6 |
| R | 116 | 54 | 90% | 81% (194 words) | 86% | 7 or 8 (four words tie for 10th) |
| L | 79 | 39 | 86% | 81% (136 words) | 84% | 5 |
| K | 48 | 26 | 91% | 83% (136 words) | 87% | 6 |
| TH | 53 | 18 | 79% | 71% (25 words) | 73% | 9 |
So the pictured words that are on the CDI are said by more 2½-year-olds than the typical CDI word with the same sound, and by more than the words on our own high-frequency lists, which add small words toddlers say less often, such as so (38%) and us (34%). Most of the 10 most-said words with each sound are nouns, such as car, juice, ball and cookie, and the picture lists have 5 to 9 of them.
The other half is unknown. 199 of the 372 picture-list words (53%) aren’t on the CDI at all, including saxophone, faucet, flying saucer, llama, flamingo, koala, mammoth, panther and thermometer. The CDI gives no data for them either way: not being on a 680-word toddler checklist doesn’t make a word rare.
Five pairs that show the gap
- “Seal” vs “see” and “so.” “Seal” is on our S and L picture lists and isn’t on the CDI. “See” is said by 88% of 2½-year-olds and “so” by 38%; both are Dolch words, as is “said”. None of the three is on our S lists.
- “Sofa” (42%) vs “sit” (92%). Both start with S. Only “sofa” is on our S lists.
- “Nurse” (39%) vs “more” (92%). “Nurse” is on our R lists; “more” isn’t, though more than 9 in 10 2½-year-olds say it.
- “Llama” and “flamingo” vs “please” (98%). Both animals are on our L picture lists and neither is on the CDI; “please” is on neither L list.
- “Raccoon” and “koala” vs “thank you” (98%). Both animals are on our K picture lists and neither is on the CDI; “thank you” is the second most-said K word on the CDI.
How we counted
- The picture lists. For each sound, the word tables for that sound on our articulation word lists (not the phrases and sentences under them), plus that sound’s flash card deck: S, R, L, K and TH, as published in November 2026. Each word is counted once per sound; a plural card is dropped when its singular is already listed (socks, sock). A word on two sounds’ lists, such as seal (S and L), counts for both.
- The high-frequency lists. Every word in the word tables and blend rows of our S, R, L, K and TH high-frequency lists, leaving out the phrases and the blank “my words” sheet.
- The CDI percentages. Wordbank’s per-item data for American English Words and Sentences at 30 months (Frank et al., 2021). Where the CDI lists a word twice (water, drink), the higher share is used. Items with an asterisk, which parents can tick for the family’s own word (grandma*, mommy*), and the name items are left out. A list word matches the same word on the CDI, singular or plural (key, keys), and the TH lists’ “thank” counts as the CDI’s “thank you”.
- Dolch. Dolch’s service words (Dolch, 1936), from the list as printed on Wikipedia in November 2026.
- Which words have the sound. Judged by sound, not spelling, with the first pronunciation in the CMU Pronouncing Dictionary (CMUdict), General American English. R includes vocalic R (car, bird) and R blends; TH includes voiced and voiceless TH. The CDI’s “close” is left out of S, because it means the verb, which ends in Z.
- Medians use only the words that are on the CDI. The “10 most-said” are the 10 CDI words with the sound that the most 2½-year-olds say; where words tie for 10th place, the table gives the range.
Limits
- Our lists only. This is an audit of our own picture-based lists. It says nothing about any other publisher’s materials.
- Toddler data, older children. The CDI is a 680-word parent-report checklist, and these figures are for 30-month-olds; our picture lists were written with children aged 4 to 8 in mind. The percentages show how early and widely children say a word, not how often a 6-year-old says it, and “not on the CDI” means no data, not a rare word.
- Dolch is about print. Dolch’s list was drawn from children’s books of the 1930s, not from speech (Dolch, 1936). It’s used here to mark small everyday words, not to measure how often children say them.
- Not independent of our newer lists. Our high-frequency lists were built from the same Wordbank and Dolch data, so their coverage shows the words fit on a list, not that the method was tested.
- Dialect. Pronunciations are General American English. ASHA notes that knowing the rules of a child’s dialect is critical in setting expectations for speech sounds (ASHA).
- No outcome claim. Research on word frequency in treatment is mixed. Frequent words led to more generalization in one small study (Morrisette & Gierut, 2002, as described in Gierut & Morrisette, 2012), frequency had no clear effect on its own in a later one (Gierut & Morrisette, 2012), and late-acquired words generalized more than early ones in another (Gierut & Morrisette, 2012). Some SLPs choose more complex targets on purpose (Storkel, 2018). The case here is functional: these are words children use all day, not words shown to speed up progress.
What this means for word lists
Pictures still earn their place: they work for children who can’t read yet, and the pictured nouns are mostly familiar. The gap is a missing row of small words. ASHA’s client-specific target selection includes relevance to the child and family, and in the core vocabulary approach the practice words are ones the child uses often in everyday life (ASHA). Gierut’s review recommends high-frequency words as treatment words, alongside late-acquired ones (Gierut, 2016).
- Add the small words. Yes, see, said, some, more, here, please, look, like, can, with: say them in short phrases (“see you”, “yes please”, “look at this”), where they’re easy to practice without a picture.
- Add the child’s own words. Names, pets, foods and words from school routines, chosen with the family.
- Choose shapes and positions as usual. Word frequency is one factor among several; see how to choose target words for articulation and target sound selection.
Our free high-frequency lists put these words first, sorted by position and word shape, with each word’s CDI share and Dolch mark: S words, R words, L words, K words and TH words.
Questions SLPs ask
Do articulation word lists use too many rare words?
In our own picture-based lists, mostly not. The pictured words that are on the CDI are said by a median of 86% to 91% of 2½-year-olds for S, R, L and K. What’s missing is the small words: the lists include 11 of the 134 Dolch service words with those sounds.
What are high-frequency words in articulation therapy?
Words children say often in everyday life, such as yes, see, more, please, look and with. Many are small words that can’t be pictured, so picture-based lists tend to leave them out.
Do high-frequency words make articulation therapy faster?
Not reliably. Frequent words led to more generalization in one small study but had no clear effect on their own in a later one, and late-acquired words did better in another. The reason to include them is functional: children use them all day.
Why don’t picture word lists include words like yes and said?
Because they can’t be drawn. When every word needs a picture, lists fill with concrete nouns, and small everyday words drop out.
Where do the percentages come from?
From Wordbank’s American English data for the MacArthur-Bates CDI: Words and Sentences. Each is the share of 30-month-olds whose parents reported they say the word. Wordbank’s data is shared under a CC BY 4.0 license.
Sources
- Frank, M. C., Braginsky, M., Yurovsky, D., & Marchman, V. A. (2017). Wordbank: An open repository for developmental vocabulary data. Journal of Child Language, 44(3), 677–694.
- Frank, M. C., Braginsky, M., Yurovsky, D., & Marchman, V. A. (2021). Variability and consistency in early language learning: The Wordbank project. MIT Press.
- Dolch, E. W. (1936). A basic sight vocabulary. The Elementary School Journal, 36(6), 456–460.
- Carnegie Mellon University. The CMU Pronouncing Dictionary (cmudict), version 0.7b and later updates.
- ASHA Practice Portal: Speech sound disorders (articulation and phonology).
- Morrisette, M. L., & Gierut, J. A. (2002). Lexical organization and phonological change in treatment. Journal of Speech, Language, and Hearing Research, 45(1), 143–159.
- Gierut, J. A., & Morrisette, M. L. (2012). Density, frequency and the expressive phonology of children with phonological delay. Journal of Child Language, 39(4), 804–834.
- Gierut, J. A., & Morrisette, M. L. (2012). Age of word acquisition effects in treatment of children with phonological delays. Applied Psycholinguistics, 33(1), 121–144.
- Storkel, H. L. (2018). The complexity approach to phonological treatment: How to select treatment targets. Language, Speech, and Hearing Services in Schools, 49(3), 463–481.
- Gierut, J. A. (2016). Nexus to lexis: Phonological disorders in children. Seminars in Speech and Language, 37(4), 280–290.