The Kid’s Kid(’s) Bed: Generic or Possessive? A Mandarin Insight

The recursive computational mechanism generates an infinite range of expressions. However, little is known about how different concepts interact with each other within recursive structures. The current study investigated how Mandarin-speaking children dealt with possessives and generics in recursive structures. The picture-matching task showed that Mandarin-speaking children 4 to 6 had a bias for generics in ambiguous possessive constructions in Mandarin, where the genitive maker was covert (e.g., Yuehan de baobao chuang John’s kid bed, where baobao chuang kid bed has both a generic interpretation and a referential interpretation). It was found that that Mandarin-speaking children below 6 had a non-recursive interpretation of the possessive John’s kid(’s) bed, and instead understand kid’s bed to refer generically to a type of bed. This finding suggests that semantics does not parallel syntax in the acquisition of indirect recursion, in line with the prediction of the generic-as-default hypothesis which claims that generics are the default mode of representation of ambiguous statements when the statement can be either generic or non-generic. The delayed recursive possessive interpretation suggests that the full determiner phrase is acquired later than a noun phrase modification, which is universal in all languages. We also discuss the role of the overt functional category in the acquisition of indirect recursion.


Introduction
An intriguing enquiry in acquisition literature is the question of how children ultimately reach the final state of grammar, but language-specific parameters and language experience complicate theoretical accounts. Empirical studies seek to establish an acquisition theory by answering the following questions: how the efficient grammar growth happens, whether developmental stages can be predicted, how different levels of linguistic representations (syntax, semantic, phonology, and pragmatics) interface, and the extent to which the universality of language can be explained (Roeper, 2021). In light of these questions, studies on the acquisition of recursive structures have emerged, since recursion is the core property of human language (Hauser et al., 2002). So far, recursive possessives have been investigated across languages (e.g., Limbach & Adone, 2010;Pérez-Leroux et al., 2012;Li et al., 2020;Akiko & Terue, 2018). One striking finding is that young children have conjunctive readings of recursive structures, suggesting that conjunction is the default grammar in the syntactic representation thereof, as articulated in many works (e.g., see in Roeper, 2011;Roeper & Oseki, 2018). In other words, the syntactic and semantic representations do not develop in parallel with the mapping between them (Muskens, 2001). Another finding is that the acquisition age of recursive structures across languages differs, which is explained by parametrical factors such as the overtness of morphosyntactic markers (Di Sciullo, 2015;Sevcenco et al., 2017) and by language-general factors such as the interface of different levels of linguistic representations (Kennedy, 2008;Roeper & Diego, 2020, in preparation;Yang et al., 2021). Unlike possessives, generics are used to express the mapping of form-referent qualities that are relatively essential, enduring, and timeless (Lyons, 1977). Although both possessives and generics are conceptual underpinnings, their linguistic expressions are language specific-for example, English has generic possessives and recursive referential possessives, whereas German has neither. Currently, little evidence is available regarding the interaction between possession and genericity in child grammar growth, let alone cross-linguistically. The current study provides a cross-linguistic comparison of the developmental trajectory of recursive possessives where generics are embedded. bike) where the possessive marker is overt, and in Mandarin (e.g., baobao che kid bike) where the possessive marker is covert. This linguistic difference helps attest the semantic-syntax interface and the language-specific parameters (i.e., the option of the covert and overt functional category) in the acquisition of indirect recursion. We find that children below 6-years prefer a generic reading when comprehending strings of interdependent and combined referential (or "regular") possessives (RPs) and generic possessives (GPs) (e.g., Bill's kid's bike, where kid's bike is a type-denoting RP), either in Mandarin or English. This work indicates that generics are the default mode of representation in indirect recursive structures that are ambiguous between generic and referential possessive readings.

Linguistic Forms of Possessives and Generics in English and Mandarin
Possessives and generics are used to organize knowledge of the world and guide inferences. Possessive structures provide an interpretation of specificity, while generics do not refer to particular entities but denote concepts of categorization. In light of the difference in semantic and conceptual representations of generics and referential possessives, a number of works seek to examine the mental organization and the developmental trajectory of these two cognitive modes. The following section discusses the cross-linguistic difference between possessives and generics.

Possessives in English and Mandarin
The possession relation between two entities can be expressed in two subtypes of possessives: prepositional (e.g., the wheel of the car) and non-prepositional (e.g., the car's wheel) (Merx, 2016). Both prepositional and non-prepositional possessives are available in English, but only non-prepositional possessives are used in Mandarin. In non-prepositional possessives, the sequence organization follows the form of NP1 POS NP2, where NP1 is possessor, and NP2 is possessum, and POS is the possessive marker. The current study concentrates on non-prepositional possessives in Mandarin and English.
In English, a possessive marker 's is obligatory to express non-prepositional possession in the linear order of NP1's NP2 (e.g., a kid's bike). By comparison, the non-prepositional possessive can be expressed in different forms (e.g., 1a-1c) in Mandarin, where the possessive marker in (1a) and (1b) can be replaced by the combination of a demonstrative (DEM) and a classifier (CL). The current study only concentrates on the form of (1a) and the possessive marker de. In Mandarin possessives, de is not a necessary word or morpheme to encode possessives, since it can be omitted in externalization in some circumstances. The omission of the possessive marker de leads to an ambiguity between a type-denoting interpretation and regular possessive interpretation (e.g., baobao chuang kid bed can refer to a specific bed or a kind of bed, namely one made for kids). However, possessives without the overt marker de (e.g., in the form of NP1 NP2) bear resemblance to compounds in terms of word formation. Thus, it is necessary and important to distinguish possessives with de omission from compounds in Mandarin. Consider the constraints of de omission in possessives. In cases where NP1 is a pronominal (e.g., wo (de) jiaxiang my hometown), and where NP1 and NP2 are temporal-related (e.g., zuotian shangwu last evening) or locative-related (e.g., Beijing Gugong the Forbidden City of Beijing), the possessive marker can be omitted (Li, 2005). The economical principle also licenses omission when the possessive appears more than once in a text (Lu, 2003). Omission is always licensed in proper names which are considered compounds (e.g., yuyan yanjiu suo the institution of language research) (Li, 2005). In contrast, in cases of prototype possessives where the omission of the possessive marker leads to a radical change of semantics and syntax (e.g., when NP2 denotes positions, titles etc.), omission is not licensed. Interestingly, the unchangeable relation between pronominal NP1 and NP2 (e.g., wo (de) fuqin my father) permits the optional use of the possessive marker (Cui, 1992).
Regarding the resemblance of the compound to the possessive with an omitted marker, multiple diagnostics have been proposed in the literature to distinguish them. One is that the functor de cannot be inserted between the components in a compound (Chao, 1968;Zhu, 1982). Another diagnostic is inseparability, i.e., a compound as a whole can be modified but the preceding component cannot be modified separately (Lieber & Pavol, 2009). This approach accords with the substitution test, according to which the post component in compounds cannot be substituted by the pronoun 'one' (Bauer, 1998). For instance, in English 'kid bed' is not equivalent to 'kid one', but one can say 'this kid's bed' and 'kid one' referring to the generic possessive, and in Mandarin 'baobao chuang' kid bed is not equivalent to 'baobao yi' kid one. Along the same vein, it is fair to say that 2(a) has either a generic reading (i.e., a dorm that is good and used by students) or a regular possessive reading like 2 (b) (i.e., a dorm that is used by good student(s)).
2(a). hao de xuesheng sushe good de student dorm good student dorm 2(b). hao de xuesheng de sushe good de student de dorm good student' s dorm

Generics in English and Mandarin
In English, generics can be expressed with multiple forms, including definite singulars (e.g., the bird is a warm-blooded animal), bare plurals (e.g., dinosaurs are extinct), and indefinite singulars (e.g., a cat has nice lives) (Lyons, 1977). A strand of studies claim that English has a variety of morphosyntactic, semantic, and pragmatic cues according to which the distinction between generics and non-generics can be drawn. More specifically, four types of morphosyntactic cues are well discussed, including determiners, number, tense, and aspect (see the review in Gelman & Tardif, 1998). In contrast, Mandarin shows substantial typological differences in terms of the way generics are specified. Noun phrases in Mandarin depend more on the pragmatic and discourse situation than on a grammatical marking of definiteness, tense, aspect, or number to gain generic interpretations. For example, the meaning-form mapping is hard to identify in cases where bare forms refer to both indefinite and generic statements and definite statements (Cheng & Sybesma, 1999). This linguistic difference brings out a longstanding enquiry of how children acquire generics in the absence of dedicated words or morphemes that encode them (Dahl, 1975). This enquiry poses the further question of whether linguistic differences in generic expression impact the interface of language and cognition, i.e., whether children with different generic representations show a universal pattern when acquiring complex structures that involve generics.
A classic approach to the interplay of language and cognition is the Generics-as-Default Hypothesis (henceforth, GaD) (Leslie, 2007(Leslie, , 2008, according to which generics are a default, primitive, and innate mode of representation in knowledge organization. This hypothesis makes predictions about child language acquisition and inductive inference, mentioned in Lazaridou-Chatzigoga et al. (2015).
Prediction 1. Age/Ease of Acquisition: children are expected to produce and comprehend generics with greater ease than quantifiers, and at earlier ages.
Prediction 2. Generic Over-generalization Effect: "non-generic generalizations would, from time to time, inappropriately exhibit some characteristics of generics, especially if the information-processing demands were made great enough" (Leslie, 2008).
For decades, the GaD hypothesis has been examined in the acquisition research field. It is well substantiated that children show less difficulty with generic interpretation than quantification in Mandarin (Tardif et al., 2012;Gelman et al., 2019) and in English (e.g., Hollander et al., 2002;Jönsson & Hampton, 2006;Leslie & Gelman, 2012;Gelman et al., 2015). Striking evidence comes from the home sign system of deaf children who use a similar number of generics compared with hearing children (Goldin-Meadow et al., 2005). Tardif et al. (2012) first compared the interpretation of generic noun phrase in Mandarin children and adults. They found Mandarin children aged 5 can distinguish generics from quantificational generalizations encoded by all or some. With reference to the earlier sensitivity of English-speaking children to generics around 3 years, they attributed this inconsistency of developmental trajectories to the grammatical marking difference. More importantly, cross-linguistic comparisons are needed to sharpen the status of generics. Indeed, Gelman and Tardif (1998) were the only and first to examine the use of generics in child language in English and Mandarin. By analyzing child-directed speech from caregivers interacting with their children aged 19−23 months, they found that languages showed a vast difference in the distribution of generics and non-generics, and that the generic advantage was more apparent for animal domain than artifacts, and that the frequency of non-generics showed no cross-linguistic difference. These results bring a pivot comparison to two contrasting views: the conceptual universal view, which holds that the use of generics depends on the conceptual system rather than language-specific forms of expression, and the linguistics relativity view, which holds that the tendency to produce generics should be influenced by means of linguistic expression. Further, the accessible ease of generics is also demonstrated by online processing (e.g., Prasada et al., 2008;Leslie et al., 2011;Meyer et al., 2011), ijel.ccsenet. which is p inhibit its d

The Cu
The the acquisition of indirect recursion (or the language-cognition interplay in a broader sense).

Methods
This work sought to provide first experimental evidence on the acquisition path of the ambiguous non-prepositional possessives with generics in English and in Mandarin Chinese. According to Zhou and Liang (2007), kinship relation and whole-part relation with respect to the human body are acquired earlier than other possessive relations, so we manipulated the possession relation type by only investigating human-object relation. We asked whether generic semantics is the default mode in the acquisition of non-prepositional recursive possessives in Mandarin and English, and whether the acquisition age of recursive possessives differs due to the option of morphosyntactic markers, and whether generics are intermediate in the acquisition of recursive structures across languages. Empirically, we predict first that young children tend to interpret more GPs than adults, second that children's acquisition of two-level recursive possessives follows three steps, and third that the acquisition of three-level recursive possessives will trigger the efficient computation of multiple-level recursive possessives.

Participants
We recruited one-hundred and three Chinese-speaking children aged four to ten who grew up in Gansu Province (located in Northwestern China). Twenty-two Mandarin-speaking college students served as controls and they had no exposure to linguistic training for the current research interest. All participants were not told the objective of the research. Children were split into 6 groups: . None of participants had hearing or visual impairment and all were typically developing. Children were chosen by the examiners at random to engage in the task, in order to minimize the effect of the teacher-student relationship.

Materials
The experiment in this study consisted of three tests for critical items and three fillers. Initially, ten Mandarin-speaking adults were recruited to write down 30 Chinese noun phrases that had both generic and possessive readings, as well as a noun phrase that only had a generic reading. And then another ten postgraduates majoring in Chinese linguistics were asked to rate the ambiguity of these phrases using a six-point scale (from 1(non-ambiguous) to 6 (most ambiguous)). Due to the possible influence of prosody, all test items follow syllable structure of 2+1, and control 2+2. Finally, three ambiguous items (i.e., kid's beds/cars/shoes) were chosen as critical items. Two-level and three-level possessives were critical items, and they were ambiguous and non-ambiguous (see in The experiment began with the introduction of 4 characters including John (in yellow), his kid (in yellow), and Bill (in blue) and his kid (in blue) (see in Figure 1a), all of whom owned two kinds of beds: a kid-type bed, smaller with guardrails, and a grownup-type bed, without guardrails (see in Figure 1b). Participants were asked to circle xiaoming de baobao John's kid and xiaoliang de baobao Bill's kid. Then two characters (John and Bill) and four beds were presented (see in Figure 1c), and participants were asked to comprehend the strings (4-1a) and (4-1c) and circle the referents. Finally, we presented four characters and all the beds together, and asked participants to circle the referents indicated (4-1a, b, c, d).

John's/Bill's kid's grownup's bed
In order to make sure that the decision-making of participants was not blind guess, a truth-value-based judgment task was designed. After all constructions were tested in the picture matching task, participants needed to judge whether the experimenter pointed to the correct referent expressed in Q1−Q4. If participants' answers were not in line with the decision-making in the matching task, participants needed to be re-examined with the corresponding stings.

Coding
We only c children's in Table 2 coded dif interpretat

Results
We only bed/shoes/ three criti unambiguo The descr adults. Re rather than groups sho two-level a children ac Figure 5). analyzed the /bike), the AN ical items we ous RPs (i.e., g iptive data sh egarding the a n recursive reg owed that chil and three-leve cross ages und igure 3. The pe  uctions into =0.804>0.05), results also of the same typ items produce ren preferred nterpretations a Figure 4). As f icated that ove l recursive RP s items across a n across ages a Vol. 11,No.  typologically different in many aspects, non-prepositional recursive possessives in both languages are left-branching. Thus, we argue that the same branching of recursion may nullify the disparity caused by typology, and that the acquisition difference would be caused by different morphosyntactic cues. In the syntactic representational level, the functional head is indispensable for indirect recursion, but the externalization in the sensory-motor system can be driven by the economical principle in the mapping of syntax and semantics. In other words, the functional head is an operator in syntactic representation, and the option of its overtness is determined in externalization. This account is in line with the Minimize Externalization principle in sensory-motor system (Di Sciullo, 2015), as well as the Pronounce the Minimum (Chomsky, 2013).

Conclusion and Further Research
To conclude, we first argue that the mapping of semantics and syntax is asymmetrical when conceptual underpinnings compete. In addition, although the morphosyntactic marker is indispensable in recursive computation, the option of overtness or covertness in externalization is subjected to The Economical Principle in the interface of the sensory-motor system and the intentional-conceptual system.
This research in essence asks a fundamental question: whether the acquisition challenge of indirect recursion is caused by syntactic computation, or by conceptual atoms. Our offline acquisition data sheds light on the comprehension of ambiguous structures, but online techniques, such as event-related potentials (ERPs) that features a higher temporal and spatial resolution, can provide immediate index in language processing and comprehension. One more factor that should be taken into account is prosodic cues such as stress and duration, which influence the interpretation of ambiguous structures. As a consequence, the effect of prosody on acquiring recursion should be examined across languages in order to frame a universal account for the acquisition path of indirect recursion.