Meeting this challenge will require rigorous assessment of how – and how early – an interface between language and core knowledge unfolds in infants. Questions concerning this interface have had a long history of spirited debate. But the question is no longer whether language and cognition are linked, but how this quintessentially human link emerges. How do infants begin to forge an interface between language and cognition in the first place? And what advantages does this link afford the developing infant mind? Questions like these have taken center stage in the developmental sciences. The accumulating evidence, exciting in its own right, bears directly on Spelke's provocative proposal. Spelke herself makes a start, marshaling evidence from infants as young as 9 months of age. But we now know that a language–cognition interface is already in place, at least in rudimentary form, far earlier.
Here, my goal is to shed light on the developmental origin and rapid unfolding of a language–cognition link in very young infants, to trace its increasing power and precision over the first year of development, and to underscore that it is imperative that we consider this link as a dynamic one that unfolds over the first year. To foreshadow, listening to language supports core cognitive capacities in infants as young as 3 months; at issue is whether and how this precocious language–cognition link ultimately serves as a gateway for a suite of conceptual and representational capacities that are the signatures of human cognition.
A surprisingly early link
By 3 months of age, before they can roll over in their own cribs or recognize the sound of their own names, simply listening to language supports infants’ core cognitive capacities, including object categorization and abstract rule-learning, and does so in a way that other carefully matched acoustic signals (e.g., sine-wave tone sequences, reversed speech) do not. Initially, this link is quite broad. Listening to vocalizations of nonhuman primates confers precisely the same cognitive benefit as listening to their native language (Ferry, Hespos, & Waxman, Reference Ferry, Hespos and Waxman2010, Reference Ferry, Hespos and Waxman2013); this initial link is also sufficiently broad to include signed (American Sign Language) as well as spoken language (Novack, Brentari, Goldin-Meadow, & Waxman, Reference Novack, Brentari, Goldin-Meadow and Waxman2021). But by 6 months, infants have tuned this link specifically to their native language(s) (for a review, see Perszyk & Waxman, Reference Perszyk and Waxman2018).
This precocious link between language and cognition, and its rapid tuning, reveals unique contributions of innate capacities and infants’ experience. The initially broad link cannot be built on experience alone: By 3 months, infants have gained ample exposure to English, but virtually none to nonhuman primate vocalizations or sign language. Yet these signals confer the same cognitive advantage at 3 and 4 months.
An increasingly precise and powerful link
A few months later, the consequences of listening to language become considerably more nuanced and powerful. By 7 months, even before infants say their first words, naming a set of objects (e.g., a dog, a horse, a duck) with the same, consistently applied name focuses infants’ attention on their commonalities and supports the formation of an object category (animal). Yet providing a distinct name for each object has a very different effect, focusing infants’ attention on distinctions among the objects and supporting their representation of each as a unique individual (LaTourrette, Chan, & Waxman, Reference LaTourrette, Chan and Waxman2023; LaTourrette & Waxman, Reference LaTourrette and Waxman2020; Xu, Cote, & Baker, Reference Xu, Cote and Baker2005). This provides infants, like adults, with exceptional conceptual and representational flexibility. For example, they can represent any object (e.g., the family dog) either as a unique individual (Rover) or as a member of an object category (e.g., a dog). This flexibility is supported by language: How an object is named – either as a unique individual or a member of a category – is instrumental to how we mentally represent that object. And this representation, in turn, has powerful downstream consequences, guiding their learning and reasoning about objects. Future work will be required to discover whether 7-month-old infants’ individual and object kind representations, guided by naming, are sufficiently robust to support higher order, and perhaps even combinatorial, processes.
A dynamic link, shaped by cascading effects of infants’ advances in language and cognition
Perhaps most importantly from a developmental vantage point, the language–cognition link is not a steady state. What an infant gleans from listening to language at 3 months will vary considerably from what they will glean at 7 months and later. Therefore, a strong developmental approach is required if we are to trace the unfolding of this link, one that takes into account the cascading and dynamic effects of infants linguistic and cognitive advances as they unfold. This is especially important because infants’ linguistic and cognitive capacities are certainly not fixed; their perception of language and ability to learn from it evolve dramatically over the first year. By 7 months, infants deftly identify distinct words in continuous speech and link them to distinct kinds of representations. Put differently, at this time, they begin to establish reference. By roughly their first birthdays, infants begin to combine property and kind concepts, evoked by predicates and nouns, respectively. But at 3 months, when infants do not yet even parse individual words from the continuous speech-stream, this kind of precision is far beyond reach.
How, then, might listening to language promote core cognition in infants so young? We have proposed that for such very young infants, listening to language engages systems of arousal and attention, heightening infants’ attention to the objects in their environment, and that this, in turn, facilitates downstream core cognitive processes, like object categorization and rule-learning (Woodruff Carr et al., Reference Woodruff Carr, Perszyk, Norton, Voss, Poeppel and Waxman2021).
In closing
Very young infants reveal a surprisingly precocious interface between language and cognition, one that becomes increasingly powerful and precise over the first year of life. This interface, which evolves considerably over infants’ first year, may provide a gateway for advancing core systems of knowledge and establishing the higher-order, abstract and combinatorial representations that distinguish human thought from that of our nearest evolutionary relatives.
You have
Access
Spelke's (Reference Spelke2022) sweeping proposal for core knowledge is breathtaking for its theoretical depth and empirical reach. And like any comprehensive theory, it raises new challenges. A foundational challenge for this proposal, as currently construed, is articulating with greater precision the place of language in the creation of knowledge. Spelke acknowledges the gravity of this challenge, especially because it is essential to her argument that language is the “glue” that permits human infants to weave together distinct pieces of core knowledge from distinct cognitive domains.
Meeting this challenge will require rigorous assessment of how – and how early – an interface between language and core knowledge unfolds in infants. Questions concerning this interface have had a long history of spirited debate. But the question is no longer whether language and cognition are linked, but how this quintessentially human link emerges. How do infants begin to forge an interface between language and cognition in the first place? And what advantages does this link afford the developing infant mind? Questions like these have taken center stage in the developmental sciences. The accumulating evidence, exciting in its own right, bears directly on Spelke's provocative proposal. Spelke herself makes a start, marshaling evidence from infants as young as 9 months of age. But we now know that a language–cognition interface is already in place, at least in rudimentary form, far earlier.
Here, my goal is to shed light on the developmental origin and rapid unfolding of a language–cognition link in very young infants, to trace its increasing power and precision over the first year of development, and to underscore that it is imperative that we consider this link as a dynamic one that unfolds over the first year. To foreshadow, listening to language supports core cognitive capacities in infants as young as 3 months; at issue is whether and how this precocious language–cognition link ultimately serves as a gateway for a suite of conceptual and representational capacities that are the signatures of human cognition.
A surprisingly early link
By 3 months of age, before they can roll over in their own cribs or recognize the sound of their own names, simply listening to language supports infants’ core cognitive capacities, including object categorization and abstract rule-learning, and does so in a way that other carefully matched acoustic signals (e.g., sine-wave tone sequences, reversed speech) do not. Initially, this link is quite broad. Listening to vocalizations of nonhuman primates confers precisely the same cognitive benefit as listening to their native language (Ferry, Hespos, & Waxman, Reference Ferry, Hespos and Waxman2010, Reference Ferry, Hespos and Waxman2013); this initial link is also sufficiently broad to include signed (American Sign Language) as well as spoken language (Novack, Brentari, Goldin-Meadow, & Waxman, Reference Novack, Brentari, Goldin-Meadow and Waxman2021). But by 6 months, infants have tuned this link specifically to their native language(s) (for a review, see Perszyk & Waxman, Reference Perszyk and Waxman2018).
This precocious link between language and cognition, and its rapid tuning, reveals unique contributions of innate capacities and infants’ experience. The initially broad link cannot be built on experience alone: By 3 months, infants have gained ample exposure to English, but virtually none to nonhuman primate vocalizations or sign language. Yet these signals confer the same cognitive advantage at 3 and 4 months.
An increasingly precise and powerful link
A few months later, the consequences of listening to language become considerably more nuanced and powerful. By 7 months, even before infants say their first words, naming a set of objects (e.g., a dog, a horse, a duck) with the same, consistently applied name focuses infants’ attention on their commonalities and supports the formation of an object category (animal). Yet providing a distinct name for each object has a very different effect, focusing infants’ attention on distinctions among the objects and supporting their representation of each as a unique individual (LaTourrette, Chan, & Waxman, Reference LaTourrette, Chan and Waxman2023; LaTourrette & Waxman, Reference LaTourrette and Waxman2020; Xu, Cote, & Baker, Reference Xu, Cote and Baker2005). This provides infants, like adults, with exceptional conceptual and representational flexibility. For example, they can represent any object (e.g., the family dog) either as a unique individual (Rover) or as a member of an object category (e.g., a dog). This flexibility is supported by language: How an object is named – either as a unique individual or a member of a category – is instrumental to how we mentally represent that object. And this representation, in turn, has powerful downstream consequences, guiding their learning and reasoning about objects. Future work will be required to discover whether 7-month-old infants’ individual and object kind representations, guided by naming, are sufficiently robust to support higher order, and perhaps even combinatorial, processes.
A dynamic link, shaped by cascading effects of infants’ advances in language and cognition
Perhaps most importantly from a developmental vantage point, the language–cognition link is not a steady state. What an infant gleans from listening to language at 3 months will vary considerably from what they will glean at 7 months and later. Therefore, a strong developmental approach is required if we are to trace the unfolding of this link, one that takes into account the cascading and dynamic effects of infants linguistic and cognitive advances as they unfold. This is especially important because infants’ linguistic and cognitive capacities are certainly not fixed; their perception of language and ability to learn from it evolve dramatically over the first year. By 7 months, infants deftly identify distinct words in continuous speech and link them to distinct kinds of representations. Put differently, at this time, they begin to establish reference. By roughly their first birthdays, infants begin to combine property and kind concepts, evoked by predicates and nouns, respectively. But at 3 months, when infants do not yet even parse individual words from the continuous speech-stream, this kind of precision is far beyond reach.
How, then, might listening to language promote core cognition in infants so young? We have proposed that for such very young infants, listening to language engages systems of arousal and attention, heightening infants’ attention to the objects in their environment, and that this, in turn, facilitates downstream core cognitive processes, like object categorization and rule-learning (Woodruff Carr et al., Reference Woodruff Carr, Perszyk, Norton, Voss, Poeppel and Waxman2021).
In closing
Very young infants reveal a surprisingly precocious interface between language and cognition, one that becomes increasingly powerful and precise over the first year of life. This interface, which evolves considerably over infants’ first year, may provide a gateway for advancing core systems of knowledge and establishing the higher-order, abstract and combinatorial representations that distinguish human thought from that of our nearest evolutionary relatives.
Financial support
This work was supported by the National Institutes of Health (R01 HD088310) and Northwestern's Institute for Policy Research.
Competing interests
None.