Elsevier

Language Sciences

Volume 83, January 2021, 101316
Language Sciences

Hierarchical clause structure as a tool for cognitive advances in early childhood

https://doi.org/10.1016/j.langsci.2020.101316Get rights and content

Highlights

  • Syntactic development in children between the ages of 2–5 provides a representational tool for stage-like cognitive advancements of early childhood.

  • Layers of sentence structure emerge in a universal progression, beginning from the small clause, progressing to the development of transitivity, finiteness and full clause.

  • Each emerging layer of syntactic hierarchy augments non-verbal representations allowing children to go beyond sensory-perceptual and affective systems in reasoning about psychologically distant phenomena.

Abstract

Humans have non-verbal means for various types of mental computations, particularly those that involve concrete objects and events accessible to direct sensory experience. Unlike other species, however, (adult) humans are also capable of reasoning about abstract or psychologically distant (i.e., beyond one's direct sensory access) concepts and events, such as hidden causes, other people's beliefs and desires, counterfactual situations, and distant past and future. While reasoning about such constructs may be possible without language, it is at best inefficient (slow, imprecise, and inconsistent).

We present an argument that the development of clausal structure in children between the ages of 2–5 provides a representational tool for stage-like cognitive advancements of early childhood, previously documented in developmental research. In our view, this influence goes beyond the conceptual domain, but involves layers of syntactic hierarchy, which augment non-verbal representations and allow children to go beyond their primary, core non-linguistic (sensory-perceptual and affective), systems in reasoning about psychologically distant phenomena. We present the Gradual Emergence theory of syntax acquisition, in which the layers of sentence structure emerge in child language in a universal progression, beginning from the least complex, Small Clause, structure, progressing first to the development of transitivity, then finiteness, and finally to full clause. We review a broad swath of evidence from developmental psychology and detail our proposal that the acquisition of hierarchical syntax in the first few years of life provides the foundation for learning about the spatially, temporally and socially distant objects and events.

Introduction

Humans are unique in their ability to transcend their immediate experiences, limited to sensory perception, self-awareness, the here and now, and to conceive of “psychologically distant” objects (Liberman and Trope, 2008), such as hidden causes, other people's beliefs and desires, and distant past and future – all beyond our direct observation. Another uniquely human characteristic is language with its hierarchical structure, compositionality and discretely infinite productivity. The nature and the directionality of the relationship between language and non-verbal cognition have engendered a longstanding and contentious debate.

The view that language affects thought (linguistic determinism), not seriously entertained as a viable theory since the writings by Whorf in 1950s, saw a resurgence in the last 2 decades. With evidence of subtle interactions between language and thought in a number of languages accumulating, several distinct positions clarifying the nature of this relationship have been proposed: from the most radical one maintaining that language is thought (“language as the language of thought”) to several much weaker versions of linguistic determinism (for a comprehensive review and arguments against the strongest version of linguistic determinism see Gentner and Goldin-Meadow, 2003; Gleitman and Papafragou, 2005; Wolff and Holms, 2011).

One mechanism via which language has been suggested to influence non-verbal cognition involves individual knowledge domains (e.g., number and space), in which linguistic representations were shown to extend or supplement non-verbal (language-independent) representations, enabling people to perform tasks more complex than those that could be performed relying solely on non-linguistic cognition (Everett, 2013; Frank et al., 2008; Gentner, 2003, 2010; Pyers and Senghas, 2009).

The “language as augmenter” (Gentner and Goldin-Meadow, 2003; Wolff and Holmes, 2011) of non-verbal cognition hypothesis is in line with the Core Knowledge theory of cognitive development (Spelke, 2003; Spelke and Kinzler, 2007), according to which humans, not unlike other species, are born with a set of core systems of representational knowledge, which allows us, starting from early infancy, to intuitively reason about sets, space/geometric relations, mechanical properties of inanimate objects, as well as agents and the intentional nature of their actions. The early emerging systems provide a basic foundation upon which later developing knowledge is built. However, this initial core knowledge is neither exceptional (with parallels in other species), nor sufficient for explaining “what makes us smart” because it is limited in a number of important ways. The uniquely human linguistic combinatorial capacity allows humans to overcome the limitations of the core knowledge systems (Spelke, 2003).

One domain in which this has been argued convincingly is number cognition: humans possess (at least) two representational systems for number concepts. One is language- (and culture-) independent, which emerges early in infancy, relies on bilateral visual-spatial cerebral networks, and is used for manipulating quantities of small magnitude or for approximation when dealing with quantities larger than 3 or 4. Another one is language-based, is used for exact calculation, involves explicit learning of mathematical language, evokes left-lateralized activation in the perisylvian language areas, and allows members of cultures that use such language to use symbolic arithmetic and other domains of symbolic mathematics (Dehaene et al., 1999; Everett, 2013; Lemer et al., 2003). This two-system theory received support in studies that showed that individuals from cultures that lack words for numbers (e.g., speakers of some Amazonian languages) and prelinguistic infants are able to compare sets of small numbers or those of larger numbers but only if the comparison could be done via approximation (i.e., relying on the language-independent number sense format). However, their capacity to perform tasks requiring exact calculations was limited by language limitations (Antell and Keating, 1983; Frank et al., 2008; Lipton and Spelke, 2003, 2004; Pica et al., 2004).

Another domain in which the enabling role of language was proposed is Theory of Mind (ToM), i.e., human ability to infer others' transitory mental states. Development of these skills is a step-wise sequential process (C. C. Peterson and Wellman, 2009; Wellman et al., 2011). Elements of the ability to infer goals and intentions of agents are present in the 1st year of life, and evidence of implicit calculations (manifested as preferential looking rather than an overt behavioral response) of others' beliefs, including false beliefs, was documented in infants at 10–15 months (Sodian, 2011).

With regards to measures requiring explicit (via a verbal or an explicit non-verbal response, such as pointing) attribution of epistemic states, it was established that the ability to reason about desires becomes fairly accurate at the age of 2, while the ability to reason about beliefs between the ages of 4 and 5 (Wellman et al., 2001). Attribution of false belief is known to be a more advanced, later developing level of social cognition than representing true belief or belief in situations when the truth is unknown. Differentiating between hidden and apparent emotion is still later developing (Wellman et al., 2011).

The evidence of gradual emergence of ToM-related skills (Leslie et al., 2004; Wellman et al., 2011) suggests that ToM includes early, limited-in-scope, abilities and later-emerging higher-level skills of much greater flexibility (C. C. Peterson and Wellman, 2009). The mechanism of development within this domain, in particular, the role of language, remains unresolved. The more advanced skills may develop out of the more basic skills with neurocognitive maturation and/or conceptual learning (Sodian, 2011) within the ToM domain itself (and/or general cognitive development). Alternatively, there may be a degree of discontinuity between the lower-level and higher-level skills, with the lower level skills being language-independent and automatic, while higher-lever skills (or at least a subset of those) being tied to language acquisition.

A hypothesis regarding language being a catalyst for ToM development (false belief attribution or FB, to be precise) (J.G. De Villiers, 2005; De Villiers, 2007; Lohmann and Tomasello, 2003) suggested that acquiring the syntax of sentence complements (i.e., subordinate clauses required as an argument by certain verbs, such as “say” and “think”) provides a representational tool for enabling a child to represent propositional attitude reports (“X believes/knows that Y”), constructions with semantics required for accurate performance on explicit FB tasks (J.G. De Villiers, 2005). The hypothesis linking explicit FB attribution to syntax acquisition is supported by data from young typically developing children and children with language delay (and normal intellect), e.g., deaf children raised by hearing parents and children with SLI (De Villiers and Pyers, 2002; P.A. De Villiers, 2005; Miller, 2004; Milligan et al., 2007; Pyers and Senghas, 2009; Rakhlin et al., 2011), who have a delay in both sentence subordination and false belief attribution, even when tested non-verbally.

We will advocate the view that language has an influence on our capacity to reason about ToM, and psychologically distant objects in general. Furthermore, we will argue that this influence goes beyond the conceptual domain and involves layers of syntactic hierarchy, which provide representational means that augment non-verbal representations and allow children to go beyond their primary, core non-linguistic (sensory-perceptual and affective) systems, providing the foundation for learning about the spatially, temporally and socially distant objects and events.

We extend the argument about the role of language in non-verbal cognition to earlier developing stages of syntax acquisition and attempt to link the acquisition of basic phrase structure with the cognitive breakthroughs reported for children between the ages of 2 and 4–5, i.e., developmental period that encompasses Brown stages: (late) 1–5, when a child's Mean Length of Utterance increases from ≈2.0 to 4.0 and all core elements of sentence structure are acquired. We propose that the acquisition of hierarchical syntax in the first few years of life provides children with the representational means that extend their non-linguistic representations and enable reasoning about psychologically distant phenomena, such as hidden causality, mental “time travel”, others' desires, beliefs, and counterfactual reasoning. We recognize that without painstaking experimental and longitudinal studies, any hypothesis remains only a conjecture. However, we believe that the key strength of our proposal is precisely that it offers clear testable (and falsifiable) predictions that we hope would generate new research helping to resolve a long-standing and difficult controversy.

Section snippets

Emergence of phrase structure: the minimal universal structure view

We adhere to the Gradual Emergence theory, similar in spirit to Radford (Radford, 1988, 1990), which holds that basic sentence structure emerges in a process, guided through the interplay of biological/maturational and experiential factors, in a sequence of universal, developmentally stable phases. We proposed a modification of this approach (Rakhlin and Progovac, 2017, parallel to Progovac's (2015) reconstruction of proto-grammar in language evolution), according to which clausal layers emerge

Syntactic hierarchy of projections: a representational tool for cognitive advances in early childhood

Humans are the only species that possesses generative capacity for hierarchical phrase structure grammar (Berwick et al., 2013; Fitch and Hauser, 2004; Friederici et al., 2017). Although there are alternative views on the role of language in thought, one view maintains that natural language is not simply a medium for externalizing thought, but a vehicle for conscious propositional thinking in a sense that “linguistic representations are the bearers of propositional thought-contents” (

Is direct evidence for the language-as-cognitive-augmenter view possible to find?

The view that we have advocated here holds that certain types of human thought, namely those that involve reasoning about entities that are causally, temporally, or socially removed from direct observation are supported and facilitated by hierarchical syntactic structures that develop, layer by layer, in most typically developing children between the ages 2 and 5. This type of thought (e.g., causal reasoning, mental time travel, representational ToM, counterfactual reasoning) is considered to

On linguistic relativity

What does our proposal imply for the neo-Whorfian view that seeks to find cognitive consequences of cross-linguistic variation with respect to grammaticalization of various semantic categories, such as tense (or gender, evidentiality, subjunctive mood, etc.) in speakers of typologically distinct languages? Showing differential non-verbal cognitive performance by speakers of languages with a corresponding linguistic contrast has been challenging because verbal mediation during a non-verbal task

Conclusions

We have presented an argument linking stage-like development of syntax with gaining representational capacity that allows children to structure information in ways that lead to important cognitive advances throughout early childhood. More specifically, we have proposed that adding increasingly complex layers to a child's grammar, layer by layer, expands children's cognitive capacity by giving them more precise and efficient ways to represent complex information than is possible relying on

Acknowledgement

The preparation of this manuscript was supported by the Wayne State University English Department’s Keal Faculty Fellowship to N. V. Rakhlin.

References (158)

  • O. Friedman et al.

    The conceptual underpinnings of pretense: pretending is not ‘behaving-as-if’

    Cognition

    (2007)
  • S.A. Gelman et al.

    Insides and essences: early understandings of the non-obvious

    Cognition

    (1991)
  • S.P. Gennari et al.

    Motion events in language and cognition

    Cognition

    (2002)
  • P. Hendriks et al.

    Production/comprehension asymmetries in language acquisition

    Lingua

    (2010)
  • P. Lavenex et al.

    Building hippocampal circuits to learn and remember: insights into the development of human memory

    Behav. Brain Res.

    (2013)
  • C. Lemer et al.

    Approximate quantities and exact number words: dissociable systems

    Neuropsychologia

    (2003)
  • A.M. Leslie et al.

    Core mechanisms in ‘theory of mind’

    Trends Cognit. Sci.

    (2004)
  • J. Lidz et al.

    What infants know about syntax but couldn’t have learned: experimental evidence for syntactic structure at 18 months

    Cognition

    (2003)
  • J. Limber

    The genesis of complex sentences

  • J.A. Lum et al.

    Working, declarative and procedural memory in specific language impairment

    Cortex

    (2012)
  • S.L. Mullally et al.

    Learning to remember: the early ontogeny of episodic memory

    Dev. Cognit. Neurosci.

    (2014)
  • A. Papafragou et al.

    Does language guide event perception? Evidence from eye movements

    Cognition

    (2008)
  • D. Adger
    (2003)
  • K. Alcock et al.

    Developmental inventories using illiterate parents as informants: Communicative Development Inventory (CDI) adaptation for two Kenyan languages

    J. of Child Lang.

    (2015)
  • S.E. Antell et al.

    Perception of Numerical Invariance in Neonates

    (1983)
  • G. Baird et al.

    Memory impairment in children with language impairment

    Dev. Med. Child Neurol.

    (2010)
  • P.J. Bauer et al.

    Parameters of remembering and forgetting in the transition from infancy to early childhood

    (2000)
  • U. Bellugi et al.

    The acquisition of language

    (1964)
  • L. Bloom et al.

    Acquisition of complementation

    J. Child Lang.

    (1989)
  • H. Borer et al.

    Bi-unique relations and the maturation of grammatical principles

    Nat. Lang. Ling. Theor.

    (1992)
  • N. Botting

    Non-verbal cognitive development and language impairment

    JCPP (J. Child Psychol. Psychiatry)

    (2005)
  • M.C. Boudreault et al.

    MacArthur Communicative Development Inventories: Validity and preliminary normative data

    La Revue d'orthophonie et d'audiologie

    (2007)
  • M. Bowerman

    Learning the structure of causative verbs: a study in the relationship of cognitive, semantic, and syntactic development

    Pap. Rep. Child Lang. Dev.

    (1974)
  • M. Bowerman

    The acquisition of complex sentences

  • M. Bowerman et al.

    The acquisition of the English causative alternation

    Crossling. Perspect. Arg. Struct. Implicat. Learnabil.

    (2008)
  • R. Brown

    A First Language: The Early Stages

    (1973)
  • R. Brown et al.

    Three processes in the child's acquisition of syntax

    Harv. Educ. Rev.

    (1964)
  • R. Brown et al.

    The acquisition of syntax

    Monogr. Soc. Res. Child Dev.

    (1964)
  • P. Carruthers

    The cognitive functions of language

    Behav. Brain Sci.

    (2002)
  • J.S. Chapin et al.

    The Family Pictures subtest of the WMS-III: relationship to verbal and visual memory following temporal lobectomy for intractable epilepsy

    J. Clin. Exp. Neuropsychol.

    (2009)
  • H. Cheung et al.

    Relative roles of general and complementation language in theory-of-mind development: evidence from Cantonese and English

    Child Dev.

    (2004)
  • N. Chomsky

    The Minimalist Program Cambridge

    (1995)
  • H. Clahsen et al.

    Functional categories in early child German

    Lang. Acqu.

    (1993)
  • D. Danks

    The Psychology of Causal Perception and Reasoning

    (2009)
  • J.G. De Villiers

    Can language acquisition give children a point of view?

  • P.A. De Villiers

    The role of language in theory-of-mind development: what deaf children tell us

  • S. Dehaene et al.

    Sources of mathematical thinking: behavioral and brain-imaging evidence

    Science

    (1999)
  • R.M. Dixon et al.
    (2006)
  • M.F. Dulay et al.

    What does the Family Pictures subtest of the Wechsler Memory Scale-III measure? Insight gained from patients evaluated for epilepsy surgery

    Clin. Neuropsychol.

    (2002)
  • C. Everett
    (2013)
  • View full text