Speech Acquisition Data

A Summary of Current Speech Acquisition Data Across 27 Languages

The word “landmark” is an Olde English word originally used to describe the expansion of the boundaries of a kingdom. Nowadays, this word is typically reserved to describe changes in legislation which will have a massive impact on our lives. I think it’s fair to use this word to describe two speech acquisition data studies that came out right before the pandemic and that have huge implications for how we do our jobs as SLPs.

New speech acquisition data published by Dr. Sharynne McLeod and Dr. Kate Crowe challenges some long-held beliefs about how and when sounds emerge. Results suggests that developmental norms are common across all languages, initial sounds develop as early as age 2, and the majority of sounds are mastered by age 5, including the American /r/.

We sat down with Dr. Sharynne McLeod of Charles Sturt University in Bathurst, Australia and had a lively conversation about the results of the cumulative research she has led and been part of. Here is a quick recap and four take-aways from the conversation.

Dr. McLeod spoke at the Bilinguistics 2023 SLP Conference on Speech Sound Disorders. People raved about her presentation. It was incredible. You can still catch the talk inside of SLP Impact. Watch this short video about the event.

Speech Acquisition Data conference

Important Take-Aways from These Decades of Research on Speech Acquisition Data

Dr. Sharynne McLeod is so easy to listen to and her humility is pervasive in everything she shares. If you are new to her work, there are some things we don’t want to skip past: Her work spans decades. She received funding and traveled extensively to the countries of many of the languages she studied. She purchased and explored oodles of evaluations tools in multiple languages to understand the normative data on which they are based. Her country is her lab: The people working with and for her are out in the rural communities, towns, and urban centers of Australia.

And what is the result? Possibly the greatest data set accumulated on speech acquisition. Here are four ideas that can have an immediate impact on your work.

1. Children Are Able to Produce Almost Every Consonant in Their Language By Age 5

Regardless of the language, most sounds are acquired by the time Kindergarten starts in the US. Think about this. Think about the amount of time it takes the average child to make it through the referral process when the teacher can’t understand them. Think about the relationship between linguistic phonology and the growth of phonology as it relates to reading.

An important distinction that she made was that occasionally there is a remnant phoneme such as /th/ that older kids still may need to focus on. However, nearly all sounds and, clearly, intelligibility is normalized by 60 months.

My take-aways: I am questioning the weightiness of the referral process in many of the schools I work in. I know that many of my students do not have any academic experience when they start Kindergarten and we need to give them a few months. But if they have been in Head Start or Pre-K or if it is spring of the Kindergarten year, we need to move the process along. I also think that I can do a great job supporting Kindergarten and Pre-K teachers in the interim.

2. Developmental Norms are Similar Across All Languages

The process of identifying which sounds to work on with a bilingual child has four parts:

  1. Test them in English and identify which sounds are being said in error
    • Check! Not a problem.
  2. Find out if those sounds exist in the home language, so we can separate second-language influence from true difficulty
    • Okay here! I can compare the sound systems of two languages.
  3. Make sure the sounds can be used in the same way/position in both languages. An easily understood example is that Spanish has B, P, M, W, G, and K. However, these don’t exist in final position. So they shouldn’t be probed as final consonants English words.
    • Not as easy but I can ask a native speaker pretty quickly or look at word charts.
  4. Make sure that the sounds come in at the same age so we don’t misdiagnose based on age of acquisition
    • This has always been the bugger! Even in our research to write Difference or Disorder we had to really dig to find speech acquisition data in the most common languages we treat. But what about Hmong, Igbo, Telugu… Good luck! Either we don’t have the data or more likely it was conducted by other professionals like child psychologists and published in the native language. This makes it difficult to find.

Here is an Arabic speech acquisition data chart that is a great example of information that is actually quite rare for English-speaking professionals.

My take-aways: These new data suggests we can work with one set of norms and therefore evaluate more accurately and quickly.

3. Initial Sounds May Develop at a Younger Age Than Expected

I’m dating myself here and admitting a fondness for beautiful speech acquisition charts on Pinterest, BUT… most images and data sets available to us start at age three. This is troubling for me when viewing the tree house chart down below. Reason being, I am wondering if early intervention programs need to be revisiting their admission criteria. I question the goals I wrote early in my career during my ECI home visits.

My take-aways: When working with and diagnosing the very young children, my goals should reflect the new norms. Also, I need to expect more out of the little guys. Yes, they are typically a lot of fun, but we need to go to work! I need to be targeting /t/, /k/, /g/, /ng/, /f/, and /y/.

4. The American English R Develops Earlier Than What Has Been Cited in the Literature and is the Same as Other Languages

Yes, /r/ is a later sound, explains Dr. McLeod. However, the collective research suggests it develops at age 5, not 6, 7, or 8.

My take-aways: How many times have I been in conversations with first-grade teachers and said “let’s wait and see.” Hmmmm…. There is part of me that is so, so relieved by these findings. I have often felt that by working with bilingual populations I have had split personalities.

“Here is me working in Spanish and this is what I focus on.”

“Here is me working in English and this is what I focus on.”

“Here is me working in Arabic/Mandarin/… and this is what I focus on.”

But I knew in my heart that the Spanish-speaking kids were pulling off /r/ at five and six (per Spanish-norms) and even the trilled /rr/! Yes, the American English R(s) are complex. But three more years needed to master it? First graders beware. I’m coming for you.

American Speech Acquisition Data

There was a point during our conversation where Dr. McLeod shared that American developmental norms are not necessarily published in journals but are instead published in test manuals. This led to their follow-up study in 2020 where they focused exclusively on American English, redrawing the lines as to when the English /r/ is expected to come in.

Several poster downloads were created to easily display the data and are available on the Charles Sturt University Speech Acquisition page. Comparing the international data set and the English data sets, the outlying /th/ sound is the major difference between languages. It should also be noted that most US norms begin at age three and the compilation of the numerous speech acquisition data studies showed that the first sounds occur earlier.


The two primary speech acquisition data research studies of note are the 2018 and 2020 studies that Dr. McLeod carried out with Dr. Kathryn Crowe. Additionally, this research has been going on over several years and can be found in research panels and published texts.

Crowe, Kathryn, and Sharynne McLeod. “Children’s English consonant acquisition in the United States: A review.” American Journal of Speech-Language Pathology 29.4 (2020): 2155-2169.

McLeod, Sharynne, and Kathryn Crowe. “Children’s consonant acquisition in 27 languages: A cross-linguistic review.” American journal of speech-language pathology 27.4 (2018): 1546-1571.

International Expert Panel on Multilingual Children’s Speech (2012). Multilingual children with speech sound disorders: Position paper. Bathurst, Australia: Research Institute for Professional Practice, Learning and Education (RIPPLE), Charles Sturt University.

McLeod, S. (2012). Multilingual children’s speech. Bathurst, NSW, Australia: Charles Sturt University.

Written by: Scott Prath

8 Comments on “A Summary of Current Speech Acquisition Data Across 27 Languages”

  1. Mirjam Blumenthal January 6, 2023 at 1:02 am #

    Great stuff, thank you! A resource I miss here is Speakaboo, a free tool to elicit and record speech with games, 20 different word lists in 20 languages http://Www.Speakaboo.io for more information. On the website you can find lots of information on speech acquisition in 20 languages. Example Polish info

    • January 6, 2023 at 11:39 am #

      Thanks for sharing Mirjam.

  2. January 6, 2023 at 7:57 am #

    Thank you for this thought provoking read! As someone who has been an SLP for over 40 years, I have to agree with the majority of what you reported. I look forward to reviewing the supporting studies in the future!

    • January 6, 2023 at 11:36 am #

      Thanks for your comment Cynthia. 40 years is a lot of experience! I always felt better about diagnosis in Spanish because the results matched what I was seeing in the classroom. This information aligns English norms with Spanish norms with a couple exceptions. It will be interesting to reassess qualifications with this data.

  3. January 6, 2023 at 8:41 am #

    Scott, thank you for this helpful information. I’m left with one important question about McLeod’s research. It makes sense to me that most students are developing these sounds by the ages listed. However, do we know what percentage of children are developing them by this age? Is it a simple majority? 70%? 90%? 95%? If I am going to label a student with a speech sound impairment, I don’t want to know that they are just “below average” but that they in the bottom 5-10%.

    I’m left with this doubt because, for example, kindergarten teachers in my elementary school who have been working for decades state that a third to a quarter of their entering kindergarten students consistently do not have their /r/ sound. Thanks for your insights!

    • January 6, 2023 at 11:32 am #

      Hi Marcia,
      Your thoughts match what was summed up really well in the ASHA article that “One particular finding that was discussed frequently by SLPs…was that English-speaking children’s reported age of acquisition of “r” for the 90%–100% criteria was 5;0–5;11 (McLeod & Crowe, 2018), younger than the age reported by Smit et al. (1990; 8;0–8;11). While English consonants share many features across the world, the consonant “r” does differ across countries.”

      This led to the follow up study that looked specifically at American productions. The percentage that they are discussing was “90%-100%” accurate. They looked at 2,491 citations and didn’t include studies that didn’t have a percentage criteria. If you look at the graph under Mean Age of Acquisition of Consonant Phonemes you see that they compared US to Global productions at 50%, 75% and 90% accuracy.

      To your last point, yes, this is a lot to wrap our minds around as US professionals and that includes Kindergarten teachers. When working with them, it makes me think that I need to ask better questions about how long a child has been in school (pre-k?) and how long they have been in an English-speaking environment, not to mention this new data.

  4. January 21, 2023 at 8:08 am #

    I love this article and loved McLeod’s presentation!

    One little soapbox I have 🙂

    I work in birth-3 on an evaluation team. I probably do 150-200 evals year. I’m always baffled by statements like “initial sounds develop *as early* as age 2”

    I’m around a lot of typically developing siblings during my evaluations, and they’re producing initial sounds in words clearly much earlier than 2 years of age (14-15 months).

    • January 23, 2023 at 12:05 pm #

      Good point. It’s the ever-present conversation around “mastery.” You are right that many children speak earlier and/or provide those sounds. The research has two main considerations, one being establishing mastery and the other being when and if we step in to support (diagnose). So based on their sample size which I believe was +18,000, their averages point to 2-years-of-age and what you observe is also correct. You also make me think about how early intervention and development/learning center push vs regular daycares. I wonder if we will see norms continue to shift lower as we focus on education at a really young age. Thanks for the comment.

Leave a Reply