The role of baby-talk words in early language development

Lead Research Organisation: University of Edinburgh
Department Name: Sch of Philosophy Psychology & Language



Adults tend to adjust the way they speak when they are talking to babies and young children. Their speech becomes slower, higher in pitch, exaggerated in intonation patterns. Their sentences become shorter and simpler with a lot of repetition. This specific register of speech is commonly known as 'baby talk', 'motherese', or more technically, infant-directed speech. Researchers have long debated whether such modifications in speech play some facilitative role in children's language development. There are indeed some indications that infant-directed speech makes certain aspects of language more accessible to the learner. For instance, it has been demonstrated that infants can identify words more easily in slow-paced speech.

One puzzling feature of infant-directed speech is the existence of 'baby-talk words', words such as 'choo-choo' and 'tummy', that are specifically used with babies. If children are ultimately to learn 'train' and 'stomach', using different words to them that mean exactly the same thing does not seem to be helpful. In fact, it is not difficult to find claims made in the media that the use of words such as 'choo-choo' and 'tummy' can hinder language development. In this project, we will examine if using baby-talk words (BTWs) has any inhibitory or facilitative effects on language learning, and then offer an explanation as to why BTWs are found in infant-directed speech across languages and cultures. Our account focuses on the sound patterns that BTWs tend to have: repeated syllables (as in 'choo-choo'), lack of consonant clusters ('tummy' as opposed to 'stomach'), and recurrent sound patterns (as in the endings in 'bunny', 'doggy', 'tummy', 'daddy'). Words with these characteristics may be easier to learn, and as such, serve as a springboard for further vocabulary development. If this is true, we can expect two things. First, infants should find words that sound like real BTWs easier to learn. Second, infants exposed to more BTWs should generally undergo faster vocabulary development at least in the earliest stage.


To test the first prediction, we will carry out a series of laboratory-based experiments with infants around the age of 14 to 18 months. They will be shown video images of unfamiliar toys with made-up names. Half the names will have BTW-like features, such as repeated syllables and lack of consonant clusters (e.g., 'neenee' and 'foofoo'). The other half will have non-BTW-like features (e.g., 'proogo' and 'smoday'). By tracking the eye movement of the infants, we will be able to tell whether infants can learn the association between the toys and their names faster when the names have BTW-like features. Furthermore, by manipulating the features between the two types of names, we will be able to determine which features (such as the repetition of syllables or the lack of consonant clusters) actually make word learning easier.

To test the second prediction, we will collect data from parents on their use of real BTWs and their children's receptive and productive vocabulary at four 3-month intervals starting at 9 months. This will be done by recording the mother's speech addressed to the child and having the mother complete a questionnaire that lists a range of words that are typically understood or produced by young children. If BTWs do help infants break into the process of word learning, we should find a correlation between the amount of BTWs used by parents and the growth of vocabulary in their children.

Planned Impact

Beyond its direct academic contributions to our understanding of the role of baby-talk words in language acquisition, the project should yield outcomes that are relevant to parents as well as practitioners in early childhood development, including nursery caretakers, health visitors and speech and language therapists. As evident from a quick search on the internet, printed media and books written for parents, there is a significant amount of interest outside academia regarding the potential impact of the use of baby-talk words on children's language development. In comparison to other aspects of infant- and child-directed speech, however, the role of baby-talk words is less well understood in research. This has led non-experts and even some practitioners to make unsubstantiated claims about the potential impact of baby-talk words, including the belief that using words such as 'choo-choo' and 'tummy' in place of the adult equivalents 'train' and 'stomach' slows down language development. Our research should clarify some of such potential misunderstandings, and provide the wider public with more scientifically sound information about how children's linguistic development can or may not be influenced by the use of register-specific lexical items.


10 25 50
Description The purpose of this project was to address the following questions about so-called baby-talk words, i.e., the set of words that are specifically used for infants, such as 'choo-choo' and 'tummy': Why do many languages have baby-talk words? Why do these words have similar sound patterns across languages, in particular, repetitions (e.g., 'choo-choo', 'night-night') and recurrent ending patterns (e.g., -ie/y, as in 'mummy', 'doggy,' and 'tummy')? Is it simply because, as is often assumed, words of these shapes are easier for immature speakers to pronounce? Or is it because, as we hypothesised, infants find it easier to learn such words even when they do not produce them?
In support of this hypothesis, we discovered the following:

1. Our analysis of baby-talk words across typological different languages showed that baby-talk words (e.g., 'bunny', 'choo-choo') do not necessarily contain sound patterns that are biomechanically easier to produce than their adult counterparts (e.g., 'rabbit', 'train'). In other words, ease for pronunciation is not the sole functional motivation behind baby-talk words.
2. When we tested 9- and 18-month-olds on their ability to pick up new words from running speech and to learn new words as labels for unfamiliar objects, we found that they perform these tasks better when the words contained repeated syllables (e.g., 'neenee', 'foofoo') than when they did not (e.g., 'neefoo', 'foonee'). The sound repetition found in many baby-talk words is likely to be a reflection of these learning biases.
3. An analysis of our longitudinal naturalistic speech corpus showed that the growth of vocabulary in infants from 9 to 21 months can be predicted by the proportion of words they are exposed to with repeated sounds and disyllabic words ending in -ie/y. Our interpretation is that 9-month-olds who hear many such words get a slight head start in vocabulary learning because these words are easier to process and serve as entry points to further lexical learning.

In addition to these findings, our research project has produced a corpus of naturalistic speech spoken to and by English-learning infants at age 9, 15 and 21 months. With 47 infants recorded at 3 developmental points, this will be one of the largest single cross-sectional/longitudinal corpora documenting what language infants receive and what language they are capable for producing before the age of 2 years. We believe the corpus will be an extremely useful resource for any future research on infants' linguistic environment and linguistic development.
Exploitation Route There are several academic routes by which our findings can be taken forward.

1. Language acquisition researchers can explore the scope and range of the kind of bootstrapping in language development played by baby-talk words.
2. This project did not fully explore the potential role of speech production; that is, infants' knowledge of their articulatory capacities may have influenced their detection and learning of novel words in our experiments. This is a question that can be explored by experts on infant speech production.
3. There are some implications for research on the connection between language acquisition and language typology. Sound repetition, which our research indicates to be favoured in learning, is often _avoided_ in the vocabulary of many languages. This suggests that human language typology is not always consistent with ease of acquisition and, therefore by implication, not necessarily a product of language learning constraints.
4. For language evolution, our findings suggest that language has evolved a sub-system that only serves to aid the process of language acquisition (e.g., baby-talk words) but otherwise makes no contribution to the main system (e.g., adult vocabulary).

One of the most obvious non-academic impact our findings have is to dispel the notion frequently found in the media and child-rearing literature that the use of baby-talk words hampers language development. We hope dissemination of our work through the health and education sectors will help to correct this misinformation.
Sectors Communities and Social Services/Policy,Education,Healthcare

Description The research team has been invited to share the findings with several non-academic organisations including the Bookbug programme of the Scottish Book Trust, Children in Need Scotland, and Starcatchers. Awareness was raised among early-years practitioners and speech pathologists in attendance of the importance of infant- and child-directed speech. According to Altmetric, one of our papers reporting the key findings from the project ("Reduplicated words are easier to learn", Language Learning & Development, vol, 12, 2016) has been featured in 22 stories in 19 news outlets, including CNN news, Parent Herald, Science Daily and Medical News Today, reaching a wide range of audience.
First Year Of Impact 2015
Sector Education,Healthcare
Impact Types Cultural,Policy & public services

Title Longitudinal corpus 
Description A longitudinal corpus of spontaneous speech involving 47 infants in the Edinburgh area and their caregivers. Recordings were made when each infant was 9, 15 and 21 months old. Each datapoint consists of 90 minutes worth of audio recording saved in .wav form. In addition, vocabulary development was assessed at each data collection point by means of a parent-filled questionnaire (modified UK CDI form). 
Type Of Material Database/Collection of data 
Year Produced 2017 
Provided To Others? Yes  
Impact When transcription is completed, this would be one of the largest longitudinal corpora of infant-directed speech in terms of the number of infants involved. It will also be the first such corpus that combines naturalistic speech sample data and questionnaire-based vocabulary data from the same individuals. These unique features will provide researchers a range of new possibilities, including statistical modelling of speech/vocabulary development. 
Description Annual Bookbug Conference, Scottish Book Trust 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Professional Practitioners
Results and Impact An invited talk at an annual conference organised by the Scottish Book Trust for early years' practitioners. The talk generated questions and discussion afterwards as well as tweets and invitations to other events.
Year(s) Of Engagement Activity 2016
Description Arts from the start inspiration day: Closing the gap. Starcatchers' Workshop (Edinburgh) 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Professional Practitioners
Results and Impact Presentation and discussion on the role of baby-talk words in early development for early years' practioners.
Year(s) Of Engagement Activity 2016
Description Edinburgh Festival event 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Public/other audiences
Results and Impact A public talk event as part of the Edinburgh Fringe Festival, organised by Summerhall. Approximately 60 people from the general public attended a talk on the role of baby-talk words in early language development. The talk was followed by questions and discussion afterwards, primarily involving the effects of parental use of baby-talk words.
Year(s) Of Engagement Activity 2014
Description Interview on BBC Radio 4 show 'Word of Mouth' 
Form Of Engagement Activity A press release, press conference or response to a media enquiry/interview
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Media (as a channel to the public)
Results and Impact I was interviewed on the Radio 4 show 'Word of Mouth' about the baby-talk word project funded by the ESRC. The show has a nation-wide audience base.
Year(s) Of Engagement Activity 2013
Description Video presentation: Before First Words 
Form Of Engagement Activity Engagement focused website, blog or social media channel
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact A short video presentation about baby-talk words through an interview conducted with the PI. This was part of a web site ("Before their First Words") designed to disseminate general scientific findings about infant communication during the first year of life []. A copy of the video on YouTube has be viewed 314 times as of 13 March 2017.
Year(s) Of Engagement Activity 2015