In ‘Being pragmatic about syntactic bootstrapping’, Hacquard (Reference Hacquard2022) argues that abstract syntax is useful for word learning, but that an additional cue, pragmatics, is both necessary and available for young children during the first steps of language acquisition. She focuses on modals and attitude verbs, where the physical context seems particularly impoverished as the sole basis for deriving meanings, and thus where linguistic cues may be particularly helpful. She convincingly shows how pragmatic and syntactic cues could be combined to help young language learners learn and infer the possible meanings of attitude verbs such as “think”, “know” or “want”. She also argues that, in some circumstances, syntax and pragmatics would need to be supplemented by semantic information from context – for instance, in the case of modals such as “might”, “can”, or “must”. We agree with Hacquard on the importance of the synergies between these different cues to meaning, and wish to add two other aspects of the input that might also be used by young children in these contexts. The aspects we describe can only be noticed when one analyzes concrete examples of what children hear in their everyday lives, something which Hacquard does very often in her work (e.g., Dieuleveut, van Dooren, Cournane & Hacquard, Reference Dieuleveut, van Dooren, Cournane and Hacquard2022; Huang, White, Liao, Hacquard & Lidz, Reference Huang, White, Liao, Hacquard and Lidz2022; Yang, Reference Yang2022). Taking into account different cues for meaning would help the field go beyond current models of syntactic bootstrapping, and create an integrated picture of the synergies between different levels of linguistic information.
The first additional piece of information that might make attitude verbs and modals easier for children to learn is their connection to the person being spoken to or about – grammatical person. With the exclusion of four examples, all examples of sentences provided by Hacquard are in the third or first person (e.g., Alex denkt, dass Frankie hier wohnt. or I think it’s time for bed!). However, the more common type of sentence children hear with modals or attitude verbs seems to be the second person sentence. For example, when we examine the Nazzal Palestinian Arabic corpus in CHILDES (Nazzal, Reference Nazzal2022), we find a total of 147 mentions of second person with the verb wantFootnote 1, and only 14 mentions of the third person want, and 16 mentions of the first person want singular. Why does it matter whether the child mostly hears attitude verbs with the second and not third person? Consider the context these two different sentences might appear in: Do you want an apple? versus Do you think they will be here soon?. In the first case the speaker would potentially be presenting the child an apple, and an affirmative answer would get that apple delivered to them. In the second case, nothing concrete would happen in the world, other than the speaker maybe providing their own opinion on the matter. Thus, want and think, in the second person, are very different in terms of their consequences and, additionally, require only knowledge of one’s own thoughts and desires (not other people’s). However, contrast this with Does Alex want an apple? versus Does Alex think they will be here soon?. These examples both require the child to consider what another person might wish or think, both will be more difficult than surveying their own desires and thoughts – and both have no real-world outcome. Note that children do not have to master the morphosyntactic distinction between the second and third person to benefit from the fact that the majority of input might be in the second person. Future corpus studies should look more closely at differences between these two types of sentences – be it in the different types of meaning conveyed by them or in the different verbs they are used with.
Prosody is the second source of information that might help children learn the different types of meaning conveyed by attitude verbs and modals. Hacquard addresses important questions about how and when children can learn the differences between declaratives, interrogatives and imperative sentences and the types of meanings (or speech acts) they tend to convey: assertions, questions, or requests. She proposes that pragmatics could serve as a filter which would allow children to reason about speakers’ intentions when they hear these types of sentences, but the presence and use of prosodic information is left out of the discussion (but see a discussion about it in a recent dissertation supervised by Hacquard: Yang, Reference Yang2022). Prosodic information seems extremely relevant to the expression of pragmatic meaning across the world’s languages, as speakers tend to use prosody for many linguistic and communicative purposes. Prosody is used, for example, to highlight important information (e.g., contrasting focus, marking old vs new information), to communicate the speaker’s state of knowledge about what they are saying (e.g., uncertainty, doubt, or disbelief), and for many other purposes (e.g., Armstrong, Andreu, Esteve-Gibert & Prieto, Reference Armstrong, Andreu, Esteve-Gibert and Prieto2016; Brown & Prieto, Reference Brown, Prieto, Haugh, Kadar and Terkourafi2021; Jun, Reference Jun and Jun2005, Reference Jun2014; Ladd, Reference Ladd2008; Pierrehumbert & Hirschberg, Reference Pierrehumbert, Hirschberg, Cohen, Morgan and Pollack1990). Hacquard herself started investigating the presence of prosodic information in these types of sentences in recent studies (see e.g., Goodhue, Hacquard & Lidz, Reference Goodhue, Hacquard and Lidz2022; Zaitsu, Wehbe, Hacquard & Lidz, Reference Zaitsu, Wehbe, Hacquard and Lidz2021), but has left the role it could play in children’s interpretations out of her current analysis.
Speakers can express pragmatic meaning – for example, by changing the intonation contour of a sentence depending on whether they give an order (e.g., You must eat meat!) vs a suggestion (e.g., You may eat meat). They can also vary their prosody depending on how certain they are about the information they are giving. Note for instance that the prosody used in a sentence like Alex must eat meat, when it is conveying that it is likely that Alex eats meat, is different from the prosody used when the same sentence is used to require Alex to eat meat: Alex must eat meat!. Thus, it is likely that parents would not use the same intonation when they ask a question, give an order or make a request, and they would not use the same intonation when they are certain about what they are saying compared to when they are uncertain about the information being conveyed.
Previous studies have found that young children can rely on such prosodic information to distinguish between declaratives, questions, and assertions (see e.g., Patel & Brayton, Reference Patel and Brayton2009; Patel & Grigos, Reference Patel and Grigos2006; Zhou, Crain & Zhan, Reference Zhou, Crain and Zhan2012) and to signal and understand internal states such as beliefs, disbeliefs, desires, feelings and emotions (Armstrong, Esteve Gibert, Hübscher, Igualada & Prieto, Reference Armstrong, Esteve Gibert, Hübscher, Igualada and Prieto2018; Armstrong & Hübscher, Reference Armstrong, Hübscher, Prieto and Esteve-Gibert2018; Hübscher, Vincze & Prieto, Reference Hübscher, Vincze and Prieto2019; Krahmer & Swerts, Reference Krahmer and Swerts2005; Pronina, Hübscher, Vilà-Giménez & Prieto, Reference Pronina, Hübscher, Vilà-Giménez and Prieto2022; Swerts & Krahmer, Reference Swerts and Krahmer2005). It is therefore highly possible that prosodic information can be used as an additional cue that might help young language learners learn the different types of meaning that attitude verbs and modals can convey. However, more studies are needed to carefully investigate this question, especially when analyzing naturalistic input addressed to children, and measuring the natural prosodic cues available in their input.
For us, syntactic bootstrapping, pragmatic bootstrapping, prosodic bootstrapping and bootstrapping word learning from context (such as in the case of second-person sentences) can all be combined within a larger framework. According to this framework, young children learn by attempting to predict which word, semantic field or linguistic structure comes next (Babineau, Havron, Dautriche, de Carvalho & Christophe, Reference Babineau, Havron, Dautriche, de Carvalho and Christophe2022; Rabagliati, Gambi & Pickering, Reference Rabagliati, Gambi and Pickering2016). This larger framework views the child’s language-learning task as figuring out the intended meaning of an interlocutor using prediction. Since human communication and behavior follow specific distributional patterns – be they syntactic, pragmatic, or any other structural field – the child could use prediction to master these patterns, and adapt their predictions when they prove wrong.
While there has been an abundance of research on syntactic bootstrapping in the past decades, and research on pragmatic and prosodic bootstrapping in young children is beginning to emerge, there is very little research examining how different sources of distributional information combine in language learning, how their weights might differ, and what factors affect this weighing in children, across different languages. Previous research found that 4 – to-5-year-olds change the weight they assign to different sources of information as a function of the relative reliability of these sources of information in context (Beretti, Havron & Christophe, Reference Beretti, Havron and Christophe2020; Yurovsky, Case & Frank, Reference Yurovsky, Case and Frank2017). One other study found that 2 – to-5-year-old Turkish children use multiple distributional linguistic cues for transitivity (not only the number of NPs but also a nominal accusative marker), and suggests that Turkish children may rely less on the number of NPs than English-speaking children, given that they expect multiple cues for transitivity (Göksun, Küntay & Naigles, Reference Göksun, Küntay and Naigles2008). More cross-linguistic studies are needed to map children’s ability to combine different cues, and the factors that affect the weight of these cues to support word learning in real-life. With those in hand, we will be in a better position to judge the relative role of pragmatic cues, and how they interact with other cues present in children’s input across the world’s languages.
In ‘Being pragmatic about syntactic bootstrapping’, Hacquard (Reference Hacquard2022) argues that abstract syntax is useful for word learning, but that an additional cue, pragmatics, is both necessary and available for young children during the first steps of language acquisition. She focuses on modals and attitude verbs, where the physical context seems particularly impoverished as the sole basis for deriving meanings, and thus where linguistic cues may be particularly helpful. She convincingly shows how pragmatic and syntactic cues could be combined to help young language learners learn and infer the possible meanings of attitude verbs such as “think”, “know” or “want”. She also argues that, in some circumstances, syntax and pragmatics would need to be supplemented by semantic information from context – for instance, in the case of modals such as “might”, “can”, or “must”. We agree with Hacquard on the importance of the synergies between these different cues to meaning, and wish to add two other aspects of the input that might also be used by young children in these contexts. The aspects we describe can only be noticed when one analyzes concrete examples of what children hear in their everyday lives, something which Hacquard does very often in her work (e.g., Dieuleveut, van Dooren, Cournane & Hacquard, Reference Dieuleveut, van Dooren, Cournane and Hacquard2022; Huang, White, Liao, Hacquard & Lidz, Reference Huang, White, Liao, Hacquard and Lidz2022; Yang, Reference Yang2022). Taking into account different cues for meaning would help the field go beyond current models of syntactic bootstrapping, and create an integrated picture of the synergies between different levels of linguistic information.
The first additional piece of information that might make attitude verbs and modals easier for children to learn is their connection to the person being spoken to or about – grammatical person. With the exclusion of four examples, all examples of sentences provided by Hacquard are in the third or first person (e.g., Alex denkt, dass Frankie hier wohnt. or I think it’s time for bed!). However, the more common type of sentence children hear with modals or attitude verbs seems to be the second person sentence. For example, when we examine the Nazzal Palestinian Arabic corpus in CHILDES (Nazzal, Reference Nazzal2022), we find a total of 147 mentions of second person with the verb wantFootnote 1, and only 14 mentions of the third person want, and 16 mentions of the first person want singular. Why does it matter whether the child mostly hears attitude verbs with the second and not third person? Consider the context these two different sentences might appear in: Do you want an apple? versus Do you think they will be here soon?. In the first case the speaker would potentially be presenting the child an apple, and an affirmative answer would get that apple delivered to them. In the second case, nothing concrete would happen in the world, other than the speaker maybe providing their own opinion on the matter. Thus, want and think, in the second person, are very different in terms of their consequences and, additionally, require only knowledge of one’s own thoughts and desires (not other people’s). However, contrast this with Does Alex want an apple? versus Does Alex think they will be here soon?. These examples both require the child to consider what another person might wish or think, both will be more difficult than surveying their own desires and thoughts – and both have no real-world outcome. Note that children do not have to master the morphosyntactic distinction between the second and third person to benefit from the fact that the majority of input might be in the second person. Future corpus studies should look more closely at differences between these two types of sentences – be it in the different types of meaning conveyed by them or in the different verbs they are used with.
Prosody is the second source of information that might help children learn the different types of meaning conveyed by attitude verbs and modals. Hacquard addresses important questions about how and when children can learn the differences between declaratives, interrogatives and imperative sentences and the types of meanings (or speech acts) they tend to convey: assertions, questions, or requests. She proposes that pragmatics could serve as a filter which would allow children to reason about speakers’ intentions when they hear these types of sentences, but the presence and use of prosodic information is left out of the discussion (but see a discussion about it in a recent dissertation supervised by Hacquard: Yang, Reference Yang2022). Prosodic information seems extremely relevant to the expression of pragmatic meaning across the world’s languages, as speakers tend to use prosody for many linguistic and communicative purposes. Prosody is used, for example, to highlight important information (e.g., contrasting focus, marking old vs new information), to communicate the speaker’s state of knowledge about what they are saying (e.g., uncertainty, doubt, or disbelief), and for many other purposes (e.g., Armstrong, Andreu, Esteve-Gibert & Prieto, Reference Armstrong, Andreu, Esteve-Gibert and Prieto2016; Brown & Prieto, Reference Brown, Prieto, Haugh, Kadar and Terkourafi2021; Jun, Reference Jun and Jun2005, Reference Jun2014; Ladd, Reference Ladd2008; Pierrehumbert & Hirschberg, Reference Pierrehumbert, Hirschberg, Cohen, Morgan and Pollack1990). Hacquard herself started investigating the presence of prosodic information in these types of sentences in recent studies (see e.g., Goodhue, Hacquard & Lidz, Reference Goodhue, Hacquard and Lidz2022; Zaitsu, Wehbe, Hacquard & Lidz, Reference Zaitsu, Wehbe, Hacquard and Lidz2021), but has left the role it could play in children’s interpretations out of her current analysis.
Speakers can express pragmatic meaning – for example, by changing the intonation contour of a sentence depending on whether they give an order (e.g., You must eat meat!) vs a suggestion (e.g., You may eat meat). They can also vary their prosody depending on how certain they are about the information they are giving. Note for instance that the prosody used in a sentence like Alex must eat meat, when it is conveying that it is likely that Alex eats meat, is different from the prosody used when the same sentence is used to require Alex to eat meat: Alex must eat meat!. Thus, it is likely that parents would not use the same intonation when they ask a question, give an order or make a request, and they would not use the same intonation when they are certain about what they are saying compared to when they are uncertain about the information being conveyed.
Previous studies have found that young children can rely on such prosodic information to distinguish between declaratives, questions, and assertions (see e.g., Patel & Brayton, Reference Patel and Brayton2009; Patel & Grigos, Reference Patel and Grigos2006; Zhou, Crain & Zhan, Reference Zhou, Crain and Zhan2012) and to signal and understand internal states such as beliefs, disbeliefs, desires, feelings and emotions (Armstrong, Esteve Gibert, Hübscher, Igualada & Prieto, Reference Armstrong, Esteve Gibert, Hübscher, Igualada and Prieto2018; Armstrong & Hübscher, Reference Armstrong, Hübscher, Prieto and Esteve-Gibert2018; Hübscher, Vincze & Prieto, Reference Hübscher, Vincze and Prieto2019; Krahmer & Swerts, Reference Krahmer and Swerts2005; Pronina, Hübscher, Vilà-Giménez & Prieto, Reference Pronina, Hübscher, Vilà-Giménez and Prieto2022; Swerts & Krahmer, Reference Swerts and Krahmer2005). It is therefore highly possible that prosodic information can be used as an additional cue that might help young language learners learn the different types of meaning that attitude verbs and modals can convey. However, more studies are needed to carefully investigate this question, especially when analyzing naturalistic input addressed to children, and measuring the natural prosodic cues available in their input.
For us, syntactic bootstrapping, pragmatic bootstrapping, prosodic bootstrapping and bootstrapping word learning from context (such as in the case of second-person sentences) can all be combined within a larger framework. According to this framework, young children learn by attempting to predict which word, semantic field or linguistic structure comes next (Babineau, Havron, Dautriche, de Carvalho & Christophe, Reference Babineau, Havron, Dautriche, de Carvalho and Christophe2022; Rabagliati, Gambi & Pickering, Reference Rabagliati, Gambi and Pickering2016). This larger framework views the child’s language-learning task as figuring out the intended meaning of an interlocutor using prediction. Since human communication and behavior follow specific distributional patterns – be they syntactic, pragmatic, or any other structural field – the child could use prediction to master these patterns, and adapt their predictions when they prove wrong.
While there has been an abundance of research on syntactic bootstrapping in the past decades, and research on pragmatic and prosodic bootstrapping in young children is beginning to emerge, there is very little research examining how different sources of distributional information combine in language learning, how their weights might differ, and what factors affect this weighing in children, across different languages. Previous research found that 4 – to-5-year-olds change the weight they assign to different sources of information as a function of the relative reliability of these sources of information in context (Beretti, Havron & Christophe, Reference Beretti, Havron and Christophe2020; Yurovsky, Case & Frank, Reference Yurovsky, Case and Frank2017). One other study found that 2 – to-5-year-old Turkish children use multiple distributional linguistic cues for transitivity (not only the number of NPs but also a nominal accusative marker), and suggests that Turkish children may rely less on the number of NPs than English-speaking children, given that they expect multiple cues for transitivity (Göksun, Küntay & Naigles, Reference Göksun, Küntay and Naigles2008). More cross-linguistic studies are needed to map children’s ability to combine different cues, and the factors that affect the weight of these cues to support word learning in real-life. With those in hand, we will be in a better position to judge the relative role of pragmatic cues, and how they interact with other cues present in children’s input across the world’s languages.