2018 Friday Session C 1615

Friday, November 2, 2018 | Session C, Terrace Lounge | 4:15pm

The spontaneous eMERGEnce of recursion in child language
I. Giblin, J. Shi, P. Zhou, C. Bill, S. Crain

Two observations are widely viewed as problematic for the claim that recursion is a basic property of human language [1,2]. One is the apparent failure of one human language to implement recursion [3,4]. The second is the reported paucity of sentences with recursion in the adult input to children, and a corresponding paucity in children’s productions [5,6]. More specifically, the findings of experimental investigations are cited as evidence that children fail to produce or comprehend sentences with nominal recursion before age 6 [e.g., 7]. The present study offers evidence that, by age 4, both Mandarin- and English-speaking children are able to comprehend and produce sentences with recursive possessive phrases (nominal recursion).

A Truth Value Judgment task with an elicitation component [8] was designed to maximize the felicitous use of nominal recursion. Whenever children rejected a puppet’s false statement about events that had taken place in a story, they were asked to justify their rejections by telling the puppet “What really happened?” In addition to 4 true and 2 false filler trials, 4 false test trials were designed so a felicitous justification could be formed by embedding an additional possessive phrase inside the possessive phrase produced by the puppet (e.g., Puppet: The pirate’s biscuit was stolen. Child: No, the pirate’s frog’s biscuit was stolen). The corresponding recursive expressions in Mandarin are formed by self-embedding the modification marker DE (e.g., haidao de qingwa de binggan ‘pirate DE frog DE biscuit’).

The child participants consistently rejected the puppet’s false statements on the test trials (Mandarin 98%; English 97%). Twenty-nine of the 30 Mandarin-speaking child participants (average 4;5) produced at least 1 sentence with recursion, and 24 produced 3 or 4. Twenty- one of the 26 English-speaking children (average 4;7) produced at least 1 sentence with nominal recursion, and 16 produced 3 or 4. In total, 95 sentences with nominal recursion were elicited from Mandarin-speaking children (79% of trials) and 70 were elicited from English-speaking children (67% of trials).

A survey of parent-child interactions in CHILDES [9] revealed 107 recursive possessive phrases by caretakers; 75 of these phrases (70%) conformed to a simple format: <proper name>’s + <common noun>’s + name (e.g., Sue’s baby’s name). A previous survey of CHILDES reported that children younger than 6 do not produce or comprehend possessive genitives [6]. A survey of three Mandarin corpora revealed no examples of DE-recursion in children’s input [10]. Despite the poverty-of-the-stimulus, every English- and Mandarin- speaking child in the present study evinced understanding of sentences with nominal recursion, and over three-quarters of the child participants produced them.

To conclude, although adult speakers tend to avoid certain recursive structures, children are nonetheless able to incorporate these recursive structures into their grammars directly, without proceeding through a simpler stage. By extension, even if adult speakers of a particular language choose not to implement recursion at all, this would be unlikely to prevent learners from incorporating recursive structures into their grammars. The findings therefore support the claim that recursion is a basic property of the human language faculty.