Semantic Development in Taiwan Mandarin-Speaking Children: A Study of V-Diao

Abstract

This study investigates the semantic development of V-diao in Taiwan Mandarin-speaking children, focusing on how they acquire both literal and non-literal meanings. Three key factors influencing this acquisition—type, metaphoricality, and context—were examined. This study recruited sixty elementary school children, divided into three age groups (7, 9, and 11 years old), along with twenty graduate students serving as an adult control group. Two truth value judgment tasks were employed: the Word-in-Sentences (WISE) task, which presents sentences containing the V-diao construction, and the Word-in-Scenarios (WISC) task, which uses stories and pictures as additional aids. The results indicated that V-diao₁ was the easiest for children to comprehend, followed by V-diao₂, while V-diao₃ and V-diao₄ were more challenging. Literal meanings of V-diao were found to be easier to acquire than non-literal ones, highlighting the metaphorical effect. Additionally, participants performed better on the WISE task than the WISC task, suggesting that contextual aids like stories and pictures did not enhance performance. These findings provide insights into the semantic development of Mandarin-speaking children and the role of metaphorical and contextual factors in language acquisition.

1. Introduction

Resultative verbs are found in many languages, including Chinese and English. In Chinese, resultative verb compounds (RVCs) consist of a verb (V1) that denotes an action1, followed by another verb (V2)2 that expresses the result of that action, such as chi wan ‘eat–finish’ and sha si ‘kill–die’. For instance, in (1a), V₁, chi ‘eat’, indicates the action, while V₂, wan ‘finish,’ represents the completion of the action. Similarly, in (1b), V₁ sha ‘kill’ refers to the action, while V₂ si ‘die’ denotes the outcome of that action. Typically, V₁ is a transitive verb, and V₂ is intransitive (Li & Thompson, 1989).

(1)	a.	Ta	chi	wan	le	pingguo.
		he	eat	finish	ASP	apple
		‘He already ate an apple.’
	b.	Zhangsan	sha	si	le	Lisi.
		Zhangsan	kill	die	ASP	Lisi
		‘Zhangsan killed Lisi.’

While English does not feature resultative verb compounds in the same way, it does have resultative constructions, where the result is expressed by an adjective or prepositional phrase. Examples include “eat the plate clean” and “paint the wall red.”

Among the diverse range of RVCs in Mandarin Chinese, V-diao emerges as a particularly compelling focus of investigation due to its linguistic typicality and cognitive complexity. The verb diao (literally ‘to fall’) spans a broad semantic spectrum—ranging from concrete meanings involving a physical downward motion to more abstract interpretations such as elimination, completion, or disappearance. This semantic flexibility reflects an advanced stage of grammaticalization, marked by polysemy and high frequency in natural discourse. Notably, V-diao illustrates key cognitive–linguistic processes such as metaphorization and metonymization (Brinton & Traugott, 2005), positioning it as an ideal case for exploring the relationship between form and meaning in language development. Given these properties, V-diao serves as both a representative and theoretically rich example for investigating how Mandarin-speaking children acquire resultative constructions, particularly those involving figurative extensions. By focusing on this structure, the present study contributes to a deeper understanding of argument structure acquisition, the cognitive development of figurative language, and the broader mechanisms of semantic generalization in child language development.

Hopper and Traugott (2003) argue that the development of RVCs involves both metaphorization and metonymization, as these processes are context-dependent and lead to lexicalization and grammaticalization. Lexicalization refers to the creation of new words with meanings that cannot be predicted from their constituent parts, while grammaticalization involves lexical terms evolving into grammatical functions (Brinton & Traugott, 2005; Hopper & Traugott, 2003).

In addition to the previously mentioned examples, there are other resultative structures resembling grammaticalized RVCs in Chinese, such as V-diao. This RVC, however, remains under-researched (Liu, 2007; Lu, 2001; Pu, 2000; Tsao, 2017). For instance, diao in Archaic Chinese originally meant baidong ‘to move up and down/back and forth’, but over time, it developed meanings like ‘to keep the strap upright’ or ‘to mobilize’. Today, the core meaning of diao in V-diao is ‘to fall downward,’ as demonstrated in (2), where someone kicks a gun and it falls.

(2)	Ta	ti	diao	le	na	ba	qiang.
	he	kick	DIAO	ASP	that	CL	gun
	‘He kicked off that gun.’

Moreover, V-diao can undergo grammaticalization, taking on metaphorical meanings. For example, in (3), hei diao ‘black DIAO’ signifies a change in state, where hands turn black, rather than a literal falling.

(3)	Shou	hei	diao	le.
	hand	black	DIAO	ASP
	‘Hands became black.’

Some V-diao expressions exhibit both literal and metaphorical interpretations, showing evidence of metaphorization, a form of figurative language3. For example, zou diao can have a literal meaning, as in walking away and disappearing, or a metaphorical meaning, as in passing away, depending on the context:

(4)	a.	Ta	zou	diao	le.
		he	walk	DIAO	ASP
		‘He walked away.’
	b.	Ta	yinwei	chehuo	zou	diao	le.
		he	because	car accident	walk	DIAO	ASP
		‘He passed away due to a car accident.’

Although earlier studies have examined RVCs and the grammaticalization of resultative verbs in Mandarin, relatively little attention has been paid to how Mandarin-speaking children acquire V-diao, especially with respect to both its literal and metaphorical meanings. Most existing research has focused either on the formal properties of RVCs or on metaphor acquisition in general, without addressing how children map polysemous resultative verbs like diao onto different semantic contexts.

Metaphor comprehension has long been a central topic in language acquisition (Hsu & Chen, 2016; Levorato & Cacciari, 2002; Nippold & Fey, 1983; Ortony, 1993), with findings consistently showing that non-literal meanings are more cognitively demanding than literal ones. While several studies have investigated the role of age and context in children’s understanding of metaphors (Bloom, 2002; Colston, 2020; Hsu & Chen, 2016; Levorato & Cacciari, 2002; Nippold & Fey, 1983), few have addressed how such factors influence the acquisition of figurative extensions within specific verb constructions in Mandarin. For instance, Hsieh (2008) found that younger children particularly benefit from contextual cues when interpreting abstract meanings, which supports the view that context reduces the cognitive load of metaphor comprehension (Gibbs & Gerrig, 1989).

Building on these insights, the present study examines how Mandarin-speaking children acquire and interpret both literal and metaphorical meanings of V-diao. We hypothesize, based on the prior literature, that metaphorical extensions of V-diao pose greater cognitive difficulty than literal meanings due to their abstractness and semantic opacity. The following questions are addressed:

(1)

How do L1 Mandarin-speaking children acquire and distinguish the different meanings of V-diao?

(2)

Are non-literal meanings of V-diao more challenging to acquire than literal ones?

(3)

What role does context play in influencing the interpretation of literal and non-literal meanings of V-diao in children?

These questions aim to shed light on the developmental trajectory of figurative language comprehension in Mandarin, as well as the specific challenges posed by polysemous resultative structures. The remainder of this paper is organized as follows: Section 2 introduces the various meanings of V-diao, and Section 3 provides a review of the relevant literature on the development and comprehension of metaphorical expressions. Section 4 describes the methodology, including participants, materials, and experimental procedures; Section 5 presents and discusses the findings in relation to existing theories. Finally, Section 6 concludes with implications for research on language acquisition and pedagogical applications.

2. Meanings of V-Diao in Taiwan Mandarin

This study extends the limited body of research on the semantic development of V-diao by examining how Mandarin-speaking children differentiate between its literal and metaphorical meanings.

2.1. Type 1: V-Diao₁ ‘to Fall Downward’

Pu (2000) explained that the main definition of the term diao is to move downwards or fall. Specifically, the first type of V-diao (V-diao₁) refers to the act of falling down in a downward direction and indicates that something has been separated from a larger object. In (5a), the verb zhuang ‘bump’ is used with diao, which shows that the light from the car was hit and then became detached from the car, falling downwards. However, in (5b), the object, huai piqi which means ‘short temper,’ is an abstract entity that cannot be physically thrown away. Thus, the sentence takes on a non-literal meaning, indicating that the speaker’s sister got rid of her short temper.

(5)	a.	Lao	Wang	zhuang	diao	ni	de	che	deng	le.
		old	Wang	bump	DIAO	you	POSS	car	lamp	ASP
		‘Old Wang bumped off your car light.’
	b.	Wo	meimei	diu	diao	huai	piqi	le.
		my	sister	throw	DIAO	bad	temper	ASP
		‘My sister got rid of her short temper.’

To summarize, V-diao₁ refers to a downward movement caused by the preceding verb, and it demonstrates several semantic features such as [+movement], [+direction], [+starting point], [+endpoint], and [-negative meaning], as proposed by Liu (2007). It can also be used with different verbs such as semelfactive verbs (e.g., zhuang meaning ‘bump’) or throwing verbs (e.g., diu meaning ‘throw’) to form V-diao, which can have either a literal or non-literal meaning.

2.2. Type 2: V-Diao₂ ‘to Eliminate Something’

V-diao₂ can be interpreted both literally and non-literally, depending on the context. For example, as seen in (6a), chi diao (‘eat up’) takes a concrete object pingguo ‘apple,’ meaning that the apple is gobbled up and then removed. Therefore, chi diao in (6a) has a literal meaning. In contrast, V-diao₂ with abstract entities and certain verbs is interpreted non-literally. For instance, in (6b), pao diao ‘run away’ is interpreted metaphorically. Unlike in (6a), the subject in (6b) is an abstract entity, baogao de geshi ‘the layout of a report.’ Because the layout cannot move itself, a similar concept of movement is mapped from the source domain, namely ‘running away.’ Hence, pao diao in (6b) refers to something not being in the correct position.

(6)	a.	Didi	chi	diao	le	binggan.
		brother	eat	DIAO	ASP	cookie
		‘My brother ate the cookie.’
	b.	Ta	baogao	de	geshi	yizhi	pao	diao.
		she	report	POSS	layout	always	run	DIAO
		‘The layout of her report always ran out of the place.’

From the examples mentioned above, V-diao₂ refers to the elimination of something, and its semantic features include [-movement], [-direction], [+starting point], [-endpoint], and [-negative meaning] (Liu, 2007). V-diao₂ can be used with several types of verbs, such as consuming verbs (chi meaning ‘eat’) and moving verbs (pao meaning ‘run’).

2.3. Type 3: V-Diao₃ ‘to Complete an Event’

V-diao₃, devoid of its core meaning of ‘falling down,’ is used to highlight the termination or completion of an event (Liu, 2007; Tsao, 2017). According to Tsao (2017), this form of diao signifies the completion of an action. Liu (2007) further suggests that it undergoes both metonymization and metaphorization processes. V-diao₃ can combine with various verbs to create the V-diao construction. For instance, in (7a), phrases like ba yifu xi diao ‘wash the clothes’ and ba biao tian diao ‘fill out a form’ indicate the completion of the actions, with the literal meanings being that the clothes have been washed and the form has been filled out. Thus, diao₃ emphasizes the state of completion. In (7b), where qing diao ‘clean up’ is used with an abstract entity, the phrase means that Zhangsan helped his mother pay off the debt, signifying that the action is completed.

(7)	a.	Wo	jintian	hen	mang	a,	yao	ba	yifu	xi	diao,
		I	today	very	busy	PART	want	BA	clothes	wash	DIAO
		ba	biao	tian	diao.
		BA	form	fill	out
		‘I am very busy today. I have to wash the clothes and fill out a form.’
		(Liu, 2007, p. 137)
	b.	Zhangsan		bang	mama	qing	diao	zhaiwu.
		Zhangsan		help	mom	clean	DIAO	debt
		‘Zhangsan helped his mom pay off the debt.’

To sum up, V-diao₃ is used to emphasize the completion of an event, and its semantic features include [-movement], [-direction], [-starting point], [+endpoint], and [-negative meaning] (Liu, 2007).

2.4. Type 4: V-Diao₄ ‘to Change a State’

Compared to the previous types of V-diao, V-diao₄ is as grammaticalized as V-diao₃, undergoing both metonymization and metaphorization processes (Liu, 2007). However, V-diao₄ differs from V-diao₃ in meaning and is often used to emphasize a change in state caused by an action. Verbs compatible with V-diao₄ are predominantly transforming verbs (deadjectival verbs) (Liu, 2007; Tsao, 2017), such as hei ‘to become black,’ huang ‘to become yellow,’ huai ‘to become broken,’ and leng ‘to become cold.’ For instance, in (8a), cai hei diao ‘vegetables turn black’ literally means that the vegetables have turned black, indirectly implying that the vegetables are no longer fresh. In contrast, in (8b), where the subject is the abstract concept of a relationship, leng diao suggests that the relationship has cooled off, indicating a deterioration or awkwardness in the situation.

(8)	a.	Cai	hei	diao	jiu	bu	xinxian	le.
		vegetable	black	DIAO	then	not	fresh	ASP
		‘Vegetables won’t be fresh if they turn black.’
	b.	Changmian	leng	diao	le.
		situation	cold	DIAO	ASP
		‘The situation becomes awkward.’

However, according to Liu (2007), because diao is being grammaticalized so rapidly, it violates the rule that only verbs with negative or neutral meanings can go with V-diao₄. This shows that even verbs with positive meanings can be used with V-diao₄, such as ganyanzheng hao diao, meaning ‘get well from dry eye syndrome.’ V-diao₄ refers to a change in state and has semantic features including [-movement], [-direction], [-starting point], [+endpoint], and [±negative meaning] (Liu, 2007). Therefore, V-diao₄ can be used with verbs that have negative meanings, neutral meanings, or even positive meanings.

2.5. A Comparison of the Four Types of V-Diao

As shown in

Table 1. A comparison of the four types of V-diao4.

Type		V-Diao1	V-Diao2	V-Diao3	V-Diao4
Semantic Meanings		To Fall Downward	To Eliminate Something	To Complete an Event	To Change a State
Semantic features	[±movement]	+	−	−	−
	[±direction]	+	−	−	−
	[±starting point]	+	+	−	−
	[±endpoint]	+	−	+	+
	[±negative meaning]	−	−	−	±
Examples		zhuang diao ‘to bump’	chi diao ‘to eat’	tian diao ‘to fill out’	hei diao ‘to turn black’

, the four types of V-diao differ in their semantic features and degrees of grammaticalization. Rather than classifying them strictly as “literal” or “non-literal,” we analyze their relationships along a continuum of semantic shifts, including directionality, abstraction, and metaphorical extension.

V-diao₁ retains its original, concrete meaning of “to fall downward,” reflecting physical movement and directionality. In contrast, V-diao₂ reflects a shift from physical motion to a metaphorical sense of elimination or disappearance (e.g., chi diao ‘eat up’), retaining some aspects of an origin point but losing directionality. V-diao₃ expresses the completion of an event (e.g., tian diao ‘fill out’), further abstracted from the original motion meaning, with its endpoint features preserved but with no implication of movement or direction. V-diao₄, which conveys a change in state (e.g., hei diao ‘turn black’), involving the highest level of abstraction, incorporates evaluative or negative connotations, distinguishing it from the others. These four types can thus be placed along a continuum of semantic and cognitive complexity. V-diao₁ is the most concrete and prototypical, involving actual physical motion. V-diao₂ reflects metonymization, where the result (elimination) is inferred from the original action. V-diao₃ involves metaphorization, representing abstract completion without physical change. V-diao₄ adds evaluative meaning to the metaphorical shift, making it the most abstract and context-dependent.

In terms of semantic features, V-diao₁ uniquely encodes movement, direction, a starting point, and an endpoint. V-diao₂ shares the feature of a starting point but lacks directionality and movement. V-diao₃ and V-diao₄ retain endpoint features but show no spatial movement. Only V-diao₄ exhibits the potential for negative evaluation, further distinguishing it semantically and pragmatically.

Crucially, the metaphorical or literal status of a given V-diao expression depends not only on the structure itself but also on the preceding verb and contextual cues. As Maudslay et al. (2020) point out, metaphorical uses presuppose both abstract and concrete senses, making some V-diao expressions context-sensitive. Tsao (2017) further notes that V-diao₁ and V-diao₂ retain lexical meanings, while V-diao₃ and V-diao₄ are more fully grammaticalized, with V-diao₂ involving metonymization.

Given the progressive abstraction and semantic complexity across the four types, we propose the following grammaticalization and acquisition order: V-diao₁ > V-diao₂ > V-diao₃ > V-diao₄. This ordering is based on the number of shared features with the prototypical V-diao₁ and the increasing cognitive demands required to interpret metaphorical and evaluative meanings.

3. The Development and Comprehension of Metaphorical Expressions

The development of metaphor comprehension is a significant area of research in cognitive linguistics, developmental psychology, and pragmatics. Studies have explored how individuals acquire, interpret, and produce figurative language, emphasizing cognitive development, contextual influences, and sociocultural factors. A growing body of research highlights the intricate interplay between linguistic, cognitive, and environmental elements in metaphor acquisition, demonstrating that this process is neither linear nor purely innate.

Vosniadou (1989) provides a foundational perspective on the gradual development of metaphor comprehension in children, arguing that younger children tend to interpret metaphors literally, while older children increasingly rely on contextual cues to infer figurative meanings. Her research challenges earlier models that attributed metaphor comprehension solely to cognitive maturation, instead emphasizing the crucial role of linguistic and situational context. This idea is further reinforced by Hsu and Chen (2016), who explore Chinese children’s understanding of the figurative meanings associated with lao (“aged”). Their findings align with the Markedness Theory (Eckman, 1977), which suggests that non-literal meanings are acquired later than literal ones. They also reveal that transparent metaphors are easier to comprehend than opaque ones and that animacy plays a key role in metaphor processing.

Building on these insights, Ahrens and Gong (2021) examine how novel metaphors5—those that are unfamiliar and not yet conventionalized—are processed within discourse-level contexts. Their experimental findings indicate that when such metaphors align with existing conceptual mappings, they are integrated more efficiently than those that diverge from familiar structures. This supports theoretical claims that contextual congruency plays a critical role in facilitating metaphor comprehension and cognitive processing. Extending this discussion to metaphor production, Levorato and Cacciari (2002) explore how figurative language develops across different age groups. They find that adolescents and adults are more skilled at generating novel metaphorical expressions compared to younger children. Moreover, their study highlights that metaphorical expression among younger participants can be significantly enhanced through the use of comparative prompts (e.g., “X is like Y”), which scaffold the figurative thinking process.

While many studies focus on metaphor comprehension and production from a cognitive perspective, some scholars engage with theoretical debates on metaphor interpretation. Genovesi (2020) challenges the traditional Gricean view that metaphorical meaning is inferred rather than explicitly stated. Instead, he argues that metaphorical content should be considered part of an utterance’s explicit meaning rather than merely an implicature. This argument aligns with contextualist theories, which emphasize that metaphorical meaning is shaped by discourse and speaker intent rather than by rigid semantic rules. A related theoretical perspective is provided by Giora (2003) in On Our Mind, where she introduces the Graded Salience Hypothesis. She posits that meaning comprehension is primarily influenced by salience—how familiar and conventional a meaning is—rather than by a strict division between literal and figurative language. Her research demonstrates that conventional metaphors, irony, and sarcasm are processed as quickly, if not more quickly, than less familiar literal meanings, challenging the Standard Pragmatic Model, which assumes that literal meanings are always accessed first.

Sociocultural and linguistic factors also play a vital role in shaping metaphor comprehension and production. Gibbs and de Macedo (2010) introduce the embodied cognition perspective, suggesting that metaphorical thinking is deeply rooted in sensorimotor experiences rather than existing solely as abstract constructs. Gibbs and Colston (2020) extend this discussion by emphasizing the sociocultural dimensions of metaphor comprehension, demonstrating that cultural and linguistic contexts significantly influence how individuals interpret figurative language. Supporting this view, Gaskin and Rundblad (2023) explore metaphor production in bilingual children, finding that verbal proficiency in both languages is a strong predictor of success in generating figurative expressions. Their findings underscore the interdependent roles of linguistic and cognitive development in mastering figurative language.

Recent research has also examined metaphor comprehension in educational contexts, focusing on how teaching strategies can enhance figurative language acquisition. Deng and Wu (2024) explore the role of multimodal teaching methods, particularly the integration of gestures, in improving students’ metaphor comprehension. Their findings indicate that synchronizing gestures with verbal explanations significantly enhances understanding, particularly when dealing with abstract concepts. Their study suggests that multimodal approaches can serve as an effective bridge for metaphor-rich content, supporting a broader pedagogical framework for metaphor instruction. The role of contextual scaffolding in language learning is further highlighted by Vosniadou (1989) and Hsu and Chen (2016), whose research underscores how providing rich contextual information helps children interpret metaphors more effectively.

Despite the valuable insights these studies provide, several limitations remain. Many investigations, including those by Vosniadou (1989) and Hsu and Chen (2016), focus predominantly on Western contexts, limiting the generalizability of their findings across diverse linguistic and cultural backgrounds. Future research should expand into cross-linguistic and cross-cultural settings to develop more comprehensive models of metaphor acquisition. Additionally, while theoretical contributions such as those of Genovesi (2020) and Giora (2003) offer compelling critiques of metaphor interpretation, they lack extensive empirical validation.

In conclusion, the growing body of research on metaphor comprehension and production highlights the multifaceted nature of figurative language acquisition. These findings consistently indicate that age, cognitive development, linguistic exposure, and sociocultural background all influence metaphor understanding. Contextual support has been shown to be a key facilitator, helping children and adults interpret figurative expressions more effectively. Sociocultural influences further shape the way individuals process and generate metaphors, reinforcing the idea that metaphorical competence is not purely a function of cognitive ability but is deeply embedded in linguistic and environmental interactions. Future studies should further investigate how these factors interact, particularly in bilingual and cross-cultural contexts, to refine our understanding of figurative language acquisition and its broader cognitive and educational implications.

4. Research Design

The primary focus of this study is to closely investigate how children interpret various types of V-diao. The experimental design aims to highlight differences in children’s comprehension across multiple contexts.

4.1. Participants

This study explored differences in understanding the metaphorical meanings of V-diao by involving elementary children and adults. Participants aged 6 to 11 were specifically recruited, as previous research highlighted the importance of this age range for observing the developmental trajectory of children’s metaphor comprehension abilities.

Table 2. Background information about the participants.

Group		Age Range	Mean Age	Number
Child groups	G1	6~7	7;3	20
	G2	8~9	9;2	20
	G3	10~11	11;3	20
Adult group		23~25	24;6	20

shows that a total of 80 participants were recruited, consisting of three child groups (ages 6–7, 8–9, 10–11) and one group of Mandarin-speaking adults.

Each child group consisted of 20 participants, totaling 60 children. Additionally, 20 Mandarin-speaking adults were recruited from a public university in northern Taiwan. These adults were graduate students with a mean age of 24.6 years. The age range of the child participants was selected to investigate the developmental progression of metaphorical language comprehension. The children were recruited from an elementary school in central Taiwan, and their selection was based on their ability to recognize and understand the characters used in the test sentences, as assessed by their homeroom teachers. This ensured that all child participants had the necessary literacy skills to fully engage with the study materials, thereby maintaining consistency in the experimental design.

4.2. Materials and Methods

To investigate the children’s interpretation of the four types of V-diao, two tasks were employed to assess participants’ understanding: a Word-in-Sentences (WISE) task and a Word-in-Scenarios (WISC) task. These true-value judgment tasks (cf. Glucksberg et al., 1982; Waldon & Degen, 2020) aimed to determine whether the participants’ performance differed between tasks and whether context influenced their ability to interpret V-diao. The test items were designed based on the four types of V-diao discussed in Section 2. Specifically, V-diao₁ refers to “falling downward,” while V-diao₂ indicates “the elimination of something.” V-diao₃ represents “the completion of an event,” and V-diao₄, which is typically combined with deadjectival verbs, describes “changing to a different state.” Furthermore, each type of V-diao could be interpreted both literally and metaphorically. Each task consisted of 16 questions: 8 sentences with literal meanings and 8 with non-literal meanings of V-diao. The distribution of these sentences was randomized to minimize order effects. To ensure comprehension, high-frequency words were selected as test items, as a familiarity with the words can aid in understanding idioms and expressions (Nippold & Rudzinski, 1993).6

The WISE task required participants to interpret sentences containing V-diao without contextual information. They were provided with sentences and asked to judge whether their interpretation matched a given explanation of the sentence’s meaning. This task assessed participants’ baseline ability to understand both the literal and non-literal meanings of V-diao solely based on linguistic information, as illustrated in

Table 3. An example of the WISE task.

Participants Saw:	Participants Heard:
	“Xiao maomi zhuang diao tade yachi.” Zhe ju hua de yisi jiushi “Yachi bujian le, ranhou wangxia” Duibudui? ‘Does the sentence “The kitten bumped down its tooth” mean “The tooth disappeared and fell down?” Yes or no?’

.

In contrast, the WISC task presented the same sentences but included pictorial scenarios and storylines to provide contextual support. This allowed researchers to evaluate whether visual and narrative cues improved participants’ comprehension, particularly for abstract or metaphorical meanings. In this task, however, the participants received storylines and images alongside the sentences, designed to assist their interpretation, as shown in

Table 4. An example in the WISC task.

Participants Saw:	Participants Heard:
	Daxiong he ta de hao pengyoumen, jueding dacheng shiguangji lai Taiwan tiyan, dan tamen dao Taiwan de shihou, dou zou diule. Suoyi tamen mei ge ren dou fasheng le yixi youqu de shiqing, xuyao xiaopengyou yi qilai panduan shiqing de dui cuo. Hao ma? ‘Daxiong and his friends decided to visit Taiwan by taking a time machine. On their way here, however, he and his friends got lost, and everyone encountered something interesting, which requires your help to judge the truth value of the task. OK?’
	Jingxiang yi dao Taiwan, renbuzhu jiu dong kan xi kan, xunzhao qita ren. ‘When he arrived in Taiwan, Jingxiang couldn’t help but look around to search for other people.’
	Zhe shihou reqing de xiaoqiang jiu pao chulaile. Dang Jingxiang kan dao zhanglang, ta jiu pao diao le. ‘At that moment, an enthusiastic cockroach greeted her, causing her to run away immediately.’
	Xiaopengyou, “Jingxiang pao diao le” jiushi “Jingxiang pao zou bujianle” Duibudui? ‘Does “Jingxiang pao diao le” mean “Jingxiang ran away and disappeared?” Yes or no?’

.

The scenarios in this task featured two storylines: one about Doraemon and the other about Teddy Bear. The first slide introduced the story, while the second and third slides displayed pictures that were viewed by the participants. The final slide presented a question.

4.3. Procedures

The participants were recruited from an elementary school with parental consent forms duly signed. The two tasks were video-recorded as part of a language activity in the classroom and were conducted with the assistance of the teachers. In the video, participants first completed the WISE task, followed by the WISC task. Before beginning, they were given two practice questions to ensure they understood the procedure. During the WISE task, participants listened to and read a sentence containing V-diao along with its meaning. They were instructed to draw a circle if they believed their interpretation matched the given meaning or a cross if it did not. A short break was provided before moving on to the WISC task. Both tasks were completed within approximately 40 min.

As previously mentioned, each task included 16 questions, with one point awarded for each correct response. No points were given for incorrect or unanswered items. The mean scores for each group across both tasks were calculated and compared, taking into account factors such as age, metaphoricality, and context. All analyses were conducted using ANOVA7 in R.

5. Results and Discussion

5.1. Type Effect of V-Diao

The mean scores of the four types of V-diao of the child groups and adult group were calculated and compared, respectively, using a one-way ANOVA, as shown in

Table 5. Type effect of V-diao within the child groups and the adult group.

	Type	V-Diao1		V-Diao2		V-Diao3		V-Diao4		p-Value
Group		Mean	SD	Mean	SD	Mean	SD	Mean	SD
Child groups		0.89	0.89	0.76	1.45	0.73	1.42	0.64	1.45	3.3 × 10−13 *
Adult group		1	0	0.99	0.22	0.99	0.31	0.99	0.31	0.517

* p-value is statistically significant.

. The scores from the WISE and WISC tasks were summed8 and then averaged9. The performances of the children showed a significant difference (p < 0.001), while no significant effect was found in the adult group (p > 0.05).

In addition, a two-way ANOVA with type effect (V-diao₁, V-diao₂, V-diao₃, V-diao₄) and age (G1, G2, G3, the adult group) as factors was conducted, which yielded a significant effect of type (F(3, 304) = 27.17, p < 0.001), a significant effect of age (F(3, 304) = 74.21, p < 0.001), and a significant effect of the interaction between these two factors (F(9, 304) = 4.57, p < 0.001), as shown in

Table 6. Two-way ANOVA on type effect and age.

	SS	Df	MS	F	p-Value
Type Effect	95.425	3	31.8	27.17	1.34 × 10−15
Age	260.57	3	86.85	74.21	4.85 × 10−36
Type EffectxAge	48.149	9	5.34	4.57	1.14 × 10−5
Residuals	355.8	304	1.17

.

Furthermore, the overall tendency of the participants’ performances on the meanings of V-diao is presented in

Table 7. Within-group performances of the participants on the meanings of V-diao.

	Type	V-Diao1		V-Diao2		V-Diao3		V-Diao4		p-Value
Group		Mean	SD	Mean	SD	Mean	SD	Mean	SD
G1		0.88	1.02	0.69	1.57	0.66	1.38	0.54	1.30	8.5 × 10−8 *
G2		0.88	0.94	0.7	1.31	0.74	1.57	0.65	1.54	0.000 *
G3		0.91	0.72	0.88	0.94	0.77	1.23	0.74	0.99	5.9 × 10−5 *
Adult group		1.00	0	0.99	0.22	0.99	0.31	0.99	0.31	0.517

* p-value is statistically significant.

.

The analysis revealed significant differences in the acquisition of the four types of V-diao among the child participants. As hypothesized, the core meaning of V-diao₁ ‘to fall’ was the easiest for the children to grasp, while more abstract meanings involving metaphorical extensions posed greater challenges. The results confirmed a hierarchical difficulty in the types of V-diao, supporting the following proposed grammaticalization order: V-diao₁ > V-diao₂ > V-diao₃ > V-diao₄.

As shown in Table 7, there were significant differences in the meanings of V-diao in terms of the performances of G1 (p < 0.001), G2 (p < 0.001), and G3 (p < 0.001), but not in those of the adults (p > 0.05). This suggests that some types of V-diao were easier for children to acquire, while some were more challenging. For adults, the difficulty of the four types of V-diao was almost the same. Moreover, all the groups performed best on V-diao₁ (G1: mean = 0.88; G2: mean: 0.88; G3: mean = 0.91; Adult: mean = 1.00), while V-diao₄ was the most challenging (G1: mean = 0.54; G2: mean: 0.65; G3: mean = 0.74; Adult: mean = 0.99), which is consistent with the hierarchical difficulty proposed in Section 2. G1 performed significantly better on V-diao₃ than on V-diao₄ (p < 0.05), indicating that V-diao₄ was much more challenging for the seven-year-olds (G1). However, the nine-year-olds (G2), the eleven-year-olds (G3), and the adults displayed a similar tendency in interpreting V-diao₃ and V-diao₄, as no significant differences between them were found, which generally matched the hierarchy assumed at first.

Regarding between-group differences, the adults performed significantly better on V-diao₁ than G1, G2, and G3 (p < 0.001, p < 0.001, p < 0.001, respectively). However, there were no significant differences among the three child groups. Children at the ages of 7, 9, and 11 performed similarly well on V-diao₁. G3 performed significantly better on V-diao₂ than G1 (p < 0.001), but not better than G2 (p > 0.05). However, the adults outperformed G1, G2, and G3 (p < 0.001, p < 0.001, p < 0.001, respectively). A similar tendency was observed for the results of V-diao₃ and V-diao₄. The adults outperformed G1, G2, and G3 on V-diao₃ (p < 0.001, p < 0.001, p < 0.001, respectively) and on V-diao₄ (p < 0.001, p < 0.001, p < 0.001, respectively). Additionally, G3 performed significantly better than G1 on V-diao₃ (p < 0.05) and V-diao₄ (p < 0.001).

Generally speaking, significant differences were found in the children’s performances (G1, G2, and G3) on the different meanings of V-diao. According to Liu (2007), V-diao₁ contains the core meaning “to fall,” while V-diao₂ “to eliminate something” undergoes the process of metonymization, and V-diao₃ “to complete an event” and V-diao₄ “to change a state” undergo both metonymization and metaphorization. In this case, the hierarchical difficulty is as follows: V-diao₁ > V-diao₂ > V-diao₃ > V-diao₄.

Previous empirical studies, such as Hsu and Chen (2016), have shown that words with a literal core meaning or those that are frequently used are the easiest for children to acquire. For instance, the literal meaning of lao, with its core definition, was found to be easier for child participants to acquire compared to other types with non-literal meanings. Furthermore, concreteness plays a crucial role in understanding and developing word meanings (Schwanenflugel, 2013). Words with concrete meanings are easier to acquire than those with abstract meanings, as abstract concepts are more challenging to process. In the present research, the semantic meaning of V-diao₁ is concrete and contains the core meaning, which shows that it is the easiest meaning to acquire among the four. This was also supported by the results of the present study, which found that the scores of V-diao₁ for G1 and G2 were significantly higher than those of V-diao₂ (G1: < 0.001; G2: < 0.001; G3: > 0.05; Adult: > 0.05). Thus, V-diao₁ was the easiest type for G1 and G2. In addition, no significant differences in V-diao₁ were found among all the child groups, indicating that G1, G2, and G3 all performed similarly well on this type.

However, V-diao₂ ‘to eliminate something’ is grammaticalized and undergoes metonymization, which involves semantic shifts and is not conventionally related to the lexical item (Paradis, 2004). Hence, it is supposed to be slightly more challenging to acquire when compared with V-diao₁. As mentioned earlier, the scores on V-diao₂ for G1 and G2 were significantly lower than those on V-diao₁. Hence, V-diao₂ was indeed more challenging than V-diao₁. Furthermore, since no significant differences were found between G1 and G2 and between G2 and G3, this shows that these three age groups had the ability to interpret V-diao₂.

Next, due to the processes of metonymization and metaphorization, the semantic meanings of V-diao₃ and V-diao₄ gradually become abstract and more grammaticalized than those of V-diao₁ and V-diao₂.

According to Heine et al. (1991), metaphors can be classified into different categories, as shown in (9), which displays a scale for the degree of metaphorical abstraction.

(9) PERSON > OBJECT > ACTIVITY > SPACE > TIME > QUALITY

V-diao₃ is used to describe the completion of an event in a time domain (ACTIVITY -> TIME), while V-diao₄ indicates the change in a state (ACTIVITY -> QUALITY). Based on the proximity of TIME and QUALITY, V-diao₃ and V-diao₄ presumably pose the same degree of difficulty for the children to acquire, since the results showed no significant differences between G1 and G2 and between G2 and G3. However, it was found that only G1, but not G2, G3, and the adult group, performed significantly differently on V-diao₃ and V-diao₄ among all the groups, indicating that V-diao₃ was obviously easier for the children at 7 to acquire than V-diao₄.

This phenomenon may be explained by looking at the test items used in the present study. All the test items related to V-diao₃ were paired with a quantifier, namely quan bu and dou, which mean ‘all’. The way that children interpreted the interaction between the quantifier and negation was different from that of adults, and this situation was rule-governed (Musolino, 1998). In addition, young children tend to map quantified noun phrases and their semantic meanings isomorphically (Musolino, 1998). For example, one of the test items in the WISC task stated that Doraemon had washed “all” the clothes and asked whether there were still some dirty clothes left, as the picture showed. Since there were still some dirty clothes left on the ground in the picture, it contradicted children’s isomorphic resolution of the quantified noun phrases, which indicated there were supposed to be no dirty clothes left due to the quantifier “all”. Hence, they tended to answer “no” and correctly answered the question, leading to a higher accuracy rate with V-diao₃ than with V-diao₄. However, this situation diminished with age, as found in Musolino (1998), who also showed that young children’s interpretation gradually moved on to the adult system.

Overall, V-diao₁ and V-diao₂ were found to be easier to acquire than V-diao₃ and V-diao₄, which posed greater challenges due to their higher levels of abstraction and grammaticalization. Furthermore, V-diao₃ and V-diao₄ were equally challenging for the participants, as both required more advanced cognitive processing. The revised hierarchical difficulty of V-diao is visually represented in Figure 1, illustrating the progressive complexity of these constructions.

The findings of this study highlight a hierarchical pattern in children’s acquisition of V-diao constructions, closely tied to the semantic complexity and degree of grammaticalization of each type. As shown in the results, V-diao₁ and V-diao₂—which are more concrete and semantically transparent—are more readily acquired by younger children, as their meanings are closely associated with observable physical experiences. In contrast, V-diao₃ and V-diao₄, which involve greater metaphorical abstraction and more advanced grammaticalization, pose greater cognitive challenges. These constructions require not only more abstract reasoning but also a heightened sensitivity to contextual and linguistic cues.

The influence of contextual information is especially evident in the interpretation of V-diao₃, where quantifiers and surrounding cues play a crucial role in meaning resolution. This underscores the importance of linguistic and environmental scaffolds for younger children who are still developing the cognitive capacity for abstract and non-literal language processing (Musolino, 1998; Paradis, 2004). These observations align with prior research emphasizing the role of concreteness, developmental readiness, and contextual support in language acquisition (Gentner, 1982; Nippold & Rudzinski, 1993). In line with this, Gibbs and Colston (2012) argue that multimodal and experience-based exposure significantly enhances children’s ability to comprehend figurative and abstract meanings—an effect particularly relevant for constructions like V-diao₃ and V-diao₄.

This study advances Tsao’s (2017) structural classification of V-diao constructions by providing empirical evidence on how children interpret and acquire them. While Tsao distinguishes the four types based on their degrees of grammaticalization, our research adds a developmental perspective, examining how semantic features—such as directionality, endpoint, and evaluative meaning—interact with children’s cognitive maturity and sensitivity to context. By linking these features to actual performance patterns, we move beyond a descriptive framework to offer insights into how children process different degrees of abstraction in language.

In conclusion, the acquisition of V-diao follows a developmental trajectory that corresponds to the semantic and cognitive complexity of each type. More concrete and physically grounded forms, like V-diao₁, are acquired earlier, while more abstract and context-dependent types, such as V-diao₄, emerge later. This pattern not only supports Tsao’s grammaticalization hierarchy but also underscores the importance of cognitive development and contextual support in figurative language acquisition.

5.2. Literal and Non-Literal Meanings of V-Diao

The second research question addressed in the present study is whether L1 Chinese children can distinguish between the literal and non-literal meanings of V-diao. In this study, literal meanings are assumed to be easier for children to acquire than non-literal meanings.

Table 8. Literal and non-literal meanings of V-diao within the child groups and the adult group.

	Type	Literal		Non-Literal		p-Value
Group		Mean	SD	Mean	SD
Child groups		0.84	2.26	0.66	2.73	8.5 × 10−9 *
Adult group		0.98	0.57	0.97	0.99	0.699

* p-value is statistically significant.

presents the mean scores across different age groups, showing that the ability to interpret non-literal meanings becomes more refined with age but does not reach adult-level proficiency until later in adolescence.

Table 9. Two-way ANOVA on metaphoricality and age.

	SS	Df	MS	F	p-Value
Metaphoricality	93.52	1	93.52	36.4	4.53 × 10−9
Age	273.65	3	91.219	35.5	9.07 × 10−20
MetaphoricalityxAge	49.45	3	16.48	6.41	0.000
Residuals	801.57	312	2.56

shows that a two-way ANOVA with metaphoricality (literal and non-literal) and age (G1, G2, G3, the adult group) as factors was conducted to examine the effects. The analysis revealed a significant effect of meaning type (F(1, 312) = 36.40, p < 0.001), a significant effect of age (F(3, 312) = 35.5, p < 0.001), and a significant interaction between the two factors (F(3, 312) = 16.48, p < 0.001).

Table 10. Overall within-group performances on the literal and non-literal meanings of V-diao.

	Type	Literal		Non-Literal		p-Value
Group		Mean	SD	Mean	SD
G1		0.80	2.94	0.61	2.66	0.001 *
G2		0.85	2.16	0.62	2.85	3.5 × 10−5 *
G3		0.88	1.30	0.77	1.84	0.002 *
Adult group		0.98	0.57	0.97	0.99	0.699

* p-value is statistically significant.

presents the acquisition tendencies of literal and non-literal meanings, as well as the mean scores and p-values for G1, G2, and G3, enabling the examination of the acquisition tendencies of all the groups.

According to Table 10, the mean scores indicate that all groups (G1, G2, G3, and Adults) found it easier to understand the literal meanings of V-diao (G1: 0.80; G2: 0.85; G3: 0.88; Adults: 0.98) than its non-literal meanings (G1: 0.61; G2: 0.62; G3: 0.77; Adults: 0.97). A one-way ANOVA analysis revealed significant differences in the performance of G1 (p < 0.05), G2 (p < 0.001), and G3 (p < 0.05), showing that the acquisition of non-literal meanings of V-diao was significantly more challenging for the children than that of literal meanings. By contrast, the performance of the adults on the literal and non-literal meanings was not significantly different (p > 0.05), suggesting that non-literal meanings were not as difficult for them to understand.

Regarding the differences between the age groups, the following comparisons of their acquisition of literal and non-literal meanings of V-diao are presented. On literal meanings, G3 did not perform significantly better than G2 (p > 0.05) or G1 (p > 0.05), and neither did G2 perform significantly better than G1 (p > 0.05). However, the adult group outperformed all the child groups significantly (p < 0.001 for all). On non-literal meanings, G3 performed significantly better than G2 (p < 0.05) and G1 (p < 0.001), and G2 did not perform significantly better than G1 (p > 0.05). Again, the adult group outperformed all the child groups significantly (p < 0.001 for all). With increasing age, it appears that children’s abilities to acquire and comprehend both literal and non-literal meanings improve. However, the mean scores for non-literal meanings for G1 and G2 were similar and not significantly different, indicating that they both had a good understanding of non-literal meanings. In conclusion, G1, G2, and G3 were at the same stage in acquiring literal meanings of V-diao. As for non-literal meanings, G1 and G2 were at the same stage, which was different from G3, indicating that older children have an easier time interpreting non-literal meanings.

The difficulty children experienced in interpreting the non-literal meanings of V-diao constructions can be better understood through the lens of metaphor processing, particularly as outlined in the Generalized Elaboration Model (GEM) proposed by Levorato and Cacciari (2002). The GEM offers a cognitive-developmental account of how metaphor comprehension evolves across five progressive phases.

In Phase 1 (birth to age 7), children primarily interpret language literally due to limited inferential abilities, semantic flexibility, and contextual integration skills. As a result, metaphors—especially those that are novel or culturally embedded—are often misunderstood or ignored. In Phase 2 (ages 8–9), children begin to utilize contextual cues to infer figurative meanings. However, their comprehension remains fragile and highly dependent on the transparency of the metaphor. Metaphors that involve cross-domain mappings or abstract associations, such as many V-diao constructions, continue to pose challenges. By Phase 3 (ages 10–12), children exhibit a significant shift in metaphor processing. They begin to treat metaphors not merely as decorative language but as structured conceptual tools that map one experiential domain onto another. This reflects their growing capacity to integrate linguistic form with conceptual structure, especially when metaphors are supported by rich context. In adulthood (Phase 5), metaphor processing becomes fluent, creative, and flexible. Adults can generate, comprehend, and reinterpret metaphors across varying degrees of conventionality, relying on well-established metaphorical schemas and entrenched cross-domain mappings. This is consistent with Conceptual Metaphor Theory (Lakoff & Johnson, 1980), which posits that metaphors are foundational to human thought.

Based on the GEM, participants in G1 and G2 should align with Phases 1 and 2, respectively. However, our findings showed no significant difference in their comprehension of V-diao metaphors, suggesting both groups may be operating within a similar developmental phase—contrary to GEM’s predicted progression. In contrast, G3 performed significantly better than G1 and G2, aligning more closely with Phase 3, yet still falling short of adult proficiency. This partially deviates from prior research indicating that children around age 9 typically outperform younger peers in figurative tasks and that 11-year-olds begin to approach adult-like competence.

Overall, our findings broadly support GEM’s developmental trajectory: younger children default to literal interpretations, older children begin to interpret metaphors with contextual support, and adults exhibit mature metaphorical competence. These developmental differences highlight that metaphor comprehension is not merely linguistic but deeply cognitive, involving the gradual integration of pragmatic reasoning, analogical mapping, and semantic abstraction—skills that emerge and refine over the course of development.

The difficulty in interpreting V-diao in this study may be attributed to the inherent complexity of the expression. Hsieh (2008) investigated various si-related constructions and found that the V+si construction (resultative verb compound, or RVC) was particularly challenging for young children due to its “cause and state change” feature. This complexity requires advanced cognitive processing, which may explain younger children’s struggles. In the current study, only a few V+si expressions were included as test items, reducing the overall task difficulty. Consequently, the children in Hsieh’s study, who were exposed to a broader and more complex range of si-related expressions, may have developed metaphorical abilities earlier than those in the present study.

The findings of this study indicate that all three child groups (G1, G2, and G3) encountered greater difficulty in interpreting the non-literal meanings of V-diao constructions compared to the adult group. Adults, in contrast, showed no significant difference between their comprehension of literal and non-literal meanings, demonstrating an advanced ability to flexibly interpret both forms. This developmental progression aligns with the principles of usage-based models of language acquisition, which emphasize that linguistic knowledge emerges from a speaker’s experience with language in use rather than from an innate universal grammar.

From a usage-based perspective (Tomasello, 2003), language learning is input-driven and probabilistic. Learners gradually build up linguistic competence by attending to the frequency and salience of their forms in communicative interactions. In this study, the difficulty experienced by children—especially with non-literal V-diao constructions—suggests that they had not yet encountered these forms frequently enough in varied and meaningful contexts to have abstracted their figurative meanings. These constructions, being semantically opaque or metaphorical, are less likely to be encountered with high frequency in child-directed speech, thus limiting opportunities for entrenchment and schematic generalization. Furthermore, the adult group’s uniform performance across literal and non-literal meanings supports the idea that repeated exposure and rich contextual experience result in greater entrenchment and more robust mental representations of both literal and figurative constructions. Adults likely had more varied input over time and had internalized not just item-specific patterns but also higher-level constructional schemas—an outcome predicted by usage-based models.

The differential performance across the age groups also reinforces the importance of contextual learning. Children appeared to rely more heavily on literal interpretations, which are often more concrete and directly tied to perceptual experience. As they accumulate linguistic input, they begin to abstract over specific instances and infer more complex figurative meanings, as predicted by item-based learning (Tomasello, 2000).

Overall, this study provides evidence for the gradual abstraction and schematization of constructions through experience. This supports the usage-based claim that grammar is not a separate module but a product of general cognitive processes acting on language input over time. The trajectory observed in the comprehension of V-diao forms—from literal to figurative—exemplifies how usage frequency, contextual richness, and age-related cognitive development interact in language learning.

Recent studies shed light on the importance of task complexity and multimodal learning. Olivero (2024) noted that task complexity influences how children’s developmental phases are assessed, suggesting that simpler tasks may mask their true abilities. In this context, multimodal learning approaches—such as pairing verbal explanations with gestures or visual aids—could help bridge the gap between Phases 2 and 3. Healey et al. (2018) found that combining verbal and non-verbal cues enhances one’s understanding of figurative language, especially for children transitioning to Phase 3.

Additionally, cultural and linguistic exposure also plays a crucial role. As Dobrovol’skij and Piirainen (2021) observed, children exposed to diverse idiomatic and metaphorical constructions in their linguistic environment tend to progress through GEM phases more quickly. This suggests that the children in the present study may benefit from increased exposure to figurative constructions like V-diao in diverse and enriched linguistic contexts.

These findings support the phase-based development of figurative competence described by GEM but highlight several factors that influence progression, including the complexity of expressions like V-diao, the design of tasks, and the role of contextual and multimodal learning. Future studies could build on this by incorporating more complex tasks and multimodal teaching methods to better understand and enhance children’s abilities to comprehend non-literal language.

5.3. Contextual Effect

Table 11. General performances on the WISE task and the WISC task within the child groups and the adult group.

	Task	WISE		WISC		p-Value
Group		Mean	SD	Mean	SD
Child groups		0.82	1.44	0.68	2.07	3.1 × 10−6 *
Adult group		0.99	0.22	0.99	0.19	0.701

* p-value is statistically significant.

presents the general performance of all the children and adults on the WISE task (without context) and the WISC task (with context). A one-way ANOVA found a significant difference among all child groups (p < 0.001), but no significant difference was found for the adult group (p > 0.05).

Table 12. Two-way ANOVA on contextual effect and age.

	SS	Df	MS	F	p-Value
Contextual Effect	52.00	1	52.00	18.90	1.8 × 10−5
Age	273.65	3	91.21	33.15	1.2 × 10−18
Contextual EffectxAge	34.23	3	11.41	4.148	0.006
Residuals	858.32	312	2.75

shows the results of a two-way ANOVA with contextual effect (WISE and WISC) and age (G1, G2, G3, the adult group) as factors. The analysis revealed a significant effect of context (F(1, 312) = 18.90, p < 0.001), a significant effect of age (F(3, 312) = 33.15, p < 0.001), and a significant interaction between the two factors (F(3, 312) = 4.14, p < 0.05).

Table 13. Overall within-group performances on the WISE and WISC tasks.

	Task	WISE		WISC		p-Value
Group		Mean	SD	Mean	SD
G1		0.71	1.08	0.68	0.94	0.484
G2		0.84	0.82	0.69	1.05	6 × 10−5 *
G3		0.92	0.55	0.73	0.88	2.3 × 10−9 *
Adult group		0.99	0.22	0.99	0.19	0.701

* p-value is statistically significant.

compares the performances on the WISE and WISC tasks, showing that context only benefited older children and adults, who likely possess the necessary cognitive maturity to integrate multiple cues.

As shown in Table 13, all child groups performed better on the WISE task than on the WISC task in terms of their mean scores (G1: WISE 0.75 > WISC 0.68; G2: WISE 0.84 > WISC 0.69; G3: WISE 0.92 > WISC 0.73), while the adults performed similarly on the two tasks (Adults: WISE 0.99 = WISC 0.99). In other words, most participants performed better when no context was provided to them. With age, the performance on both tasks improved for all participants.

Regarding within-group performance, a one-way ANOVA showed no significant differences for G1 (p > 0.05) and the adult group (p > 0.05). However, G2 and G3 performed significantly better on the WISE task than on the WISC task (G2: p < 0.001; G3: p < 0.001). This finding suggests that the tasks without or with context presented a similar degree of difficulty for G1 and the adult group in comprehending V-diao, but not for G2 and G3.

In terms of the between-group differences on the WISE task, the adult group outperformed G1 (p < 0.001), G2 (p < 0.001), and G3 (p < 0.001). G3 performed significantly better than G1 (p < 0.001) and G2 (p < 0.05), and there was also a significant difference between G1 and G2 (p < 0.001). However, on the WISC task, only the adults outperformed G1 (p < 0.001), G2 (p < 0.001), and G3 (p < 0.001). No significant differences were found between G3 and G2 (p > 0.05), between G3 and G1 (p > 0.05), or between G2 and G1 (p > 0.05). Overall, when no context was provided, the participants performed significantly better with increasing age. However, when context was provided, it did not significantly improve the children’s comprehension of V-diao.

Previous studies (Gibbs & Gerrig, 1989; Hsieh, 2008; Hsu & Chen, 2016; Ortony, 1993) have consistently argued that context enhances one’s comprehension of non-literal meanings. These studies suggest that children find it easier to grasp metaphorical meanings when presented within a contextual framework. For example, Hsieh (2008) demonstrated that contextual cues could aid children in interpreting idiomatic expressions by providing inferential scaffolding. Similarly, Ortony (1993) emphasized that rich, relatable contexts could serve as cognitive bridges, enabling younger participants to decode abstract meanings.

However, other research (Y.-C. E. Cheng & Chen, 2023; Schatz & Baldwin, 1986; Shinjo, 1986) challenges the idea that context always facilitates comprehension. Consistent with these findings, the present study revealed that G1, G2, and G3 generally performed better on tasks when no context was provided. Shinjo (1986), for example, explored the role of semantic priming in metaphor comprehension and found that while priming may activate related concepts, its effects on literal and metaphorical meanings were comparable. Furthermore, if the priming context was unrelated to the metaphor, its influence was negligible. Similarly, Schatz and Baldwin (1986) conducted a series of experiments examining the impact of context on inferring word meanings, concluding that the benefits of context were overstated. Their findings suggested that participants could not derive significant advantages from contextual information, particularly when tasked with inferring novel meanings. Y.-C. E. Cheng and Chen (2023) extended these findings by examining Mandarin frequency adverbs in both context-free and context-provided tasks. Their results highlighted that older children and adults did not benefit from contextual support, likely due to the increased complexity of processing such information. Notably, they adopted a methodological approach similar to the present study, presenting test items as sentences rather than isolated phrases, which provided a more robust comparison.

The Global Elaboration Hypothesis (Levorato et al., 2004) posits that children employ specific strategies to interpret idiomatic expressions and noun phrases in context, such as hypothesizing meanings at the sentence level and filtering irrelevant interpretations. According to this hypothesis, presenting metaphorical expressions in sentences with context should provide additional cues to facilitate comprehension. However, Y.-C. E. Cheng and Chen (2023) found that their participants struggled with the context-provided task due to its complexity, which required them to process multiple sentences and story plots within a limited time. This finding aligns with the Limited Cognitive Capacity Hypothesis (Skehan & Foster, 2001), which asserts that excessive cognitive demands can hinder performance. Therefore, Cheng’s participants performed better on context-free tasks, a result mirrored in the present study.

Contrary to earlier research emphasizing the facilitative role of context in language comprehension, our findings revealed that contextual enrichment did not consistently support children’s understanding of non-literal meanings in V-diao constructions. When analyzed through the lens of Construction Grammar (Goldberg, 1995), these results highlight the developmental limitations in children’s ability to process and integrate form–meaning pairings within dynamic discourse contexts.

In Construction Grammar, constructions are learned and represented as holistic, experience-based pairings of form and meaning, ranging from fully lexical items to abstract syntactic templates. The V-diao construction, particularly in its non-literal usage, functions as a partially schematic construction—that is, one where the meaning cannot be fully predicted by the individual lexical items or syntactic form alone. Understanding such constructions requires recognizing a constructional schema that maps specific morphosyntactic patterns onto idiomatic or metaphorical meanings. This mapping is cognitively demanding and typically acquired through frequent, context-rich exposure over time.

In our study, younger children (G1, G2, and G3) performed better on tasks without contextual support, suggesting that the addition of story-based and visual context imposed a cognitive load that exceeded their processing capacity. From a constructional perspective, this indicates that the children may have not yet entrenched the V-diao construction in its metaphorical sense. When multiple semiotic inputs—text, image, and construction—competed for attention, the children struggled to establish the intended form–meaning link. Their reliance on literal interpretations, especially when primed by visual cues, further implies that non-literal constructions are still “marked” and cognitively effortful at this stage of development.

This integration difficulty was particularly evident when context was delivered through illustrated narratives. Children frequently fixated on the literal truth conditions presented in the image, rather than evaluating whether the visual narrative supported or contradicted the constructional meaning encoded in the sentence. In Construction Grammar terms, this suggests a failure to align multimodal cues with the abstracted constructional meaning. The pictures may have reinforced default (literal) readings of the construction rather than promoting the recognition of its idiomatic or metaphorical sense. Adults, by contrast, performed equally well with or without contextual support, reflecting their highly entrenched and flexible representation of both literal and non-literal V-diao constructions. For them, context served more as a confirmatory aid than a source of interpretation. Their ability to draw on a network of related constructions allowed them to bypass the surface form and directly access the intended figurative meaning, a hallmark of constructional fluency. Interestingly, while G3 outperformed G1 and G2, their continued struggle with integrating contextual and linguistic cues indicates that they are in transition toward developing abstract, context-sensitive constructional knowledge. They may have begun to recognize the V-diao construction as a multi-functional form but lack the schematic abstraction and processing efficiency seen in adults.

Ultimately, these findings suggest that contextual information does not uniformly facilitate comprehension; rather, its usefulness depends on the learner’s ability to map it onto the relevant constructional schema. Until children internalize these mappings through repeated exposure and pattern abstraction, additional context may function less as a scaffold and more as a distraction—especially in the case of non-transparent or metaphorical constructions.

Glenberg and Langston (1992) observed that while texts and images are typically processed separately in the mind, carefully designed visuals can enhance comprehension when paired with text. However, this effect was not observed in the child participants of the current study. Building on this, L. Cheng et al. (2024) examined the development of metaphor comprehension in Chinese children aged 5–8. They identified three developmental stages: the “perception period” (age 5), when children began recognizing metaphorical language; the “development period” (ages 6–7), marked by a gradual improvement in comprehension; and the “rational decision period” (age 8), characterized by faster and more accurate understanding. This research emphasizes the important roles of age and salience in shaping metaphor comprehension within a Chinese cultural context. Similarly, Inhoff et al. (1984) investigated how context affects cognitive load during metaphor comprehension. They found that metaphors were processed more efficiently when preceded by metaphorical contexts, while literal contexts worked best for literal targets. These findings reinforce the importance of simplicity in designing effective learning environments for young children. However, while context is often viewed as a helpful aid for understanding metaphors and idioms, this study and others (e.g., Y.-C. E. Cheng & Chen, 2023; Schatz & Baldwin, 1986; Shinjo, 1986) suggest that its effects are not always positive. For younger learners, overly complex contexts may hinder rather than enhance comprehension.

In conclusion, the relationship between context and comprehension depends heavily on the cognitive development of the learner. Future research should aim to identify the ideal level of contextual support for different age groups to optimize comprehension, especially in metaphorical and idiomatic language learning.

6. Conclusions

This study examined the interactions between age and three key factors—V-diao type, metaphorical complexity, and context—highlighting age as a pivotal variable in how these elements influence the acquisition of V-diao constructions. Statistical analysis (two-way ANOVA) revealed that all three factors had significant interaction effects with age. Among them, the type of V-diao had the most pronounced influence (p = 1.14 × 10⁻⁵), followed by metaphorical complexity (p = 0.000), and then context (p = 0.006).

These findings suggest that the semantic distinctions among the different V-diao types (from V-diao₁ to V-diao₄) are particularly important in shaping children’s ability to acquire this construction. As children mature, their capacity to understand and produce metaphors tends to align with their broader conceptual development (Vosniadou, 1987), pointing to a developmental trajectory in figurative language acquisition.

In addition to the influence of age, this study confirms prior findings that literal meanings are generally easier for children to comprehend and produce compared to non-literal ones (Hsu & Chen, 2016). This reinforces the notion that age and metaphoricality exhibit language-universal trends. In contrast, the effects of V-diao type and context appear to be language-specific. Variations in semantic complexity among the V-diao types played a role in how easily children acquired each construction. Interestingly, while earlier research (e.g., Hsu & Chen, 2016) suggests that supportive contexts enhance comprehension, our results did not find contextual cues to significantly improve children’s understanding of V-diao.

The findings of this study carry important pedagogical implications for Mandarin language instruction. Since V-diao₃ and V-diao₄ were identified as the most challenging constructions for children to comprehend, we recommend introducing V-diao constructions in a developmentally sensitive sequence—starting with more concrete, literal uses, and gradually moving toward more abstract and figurative meanings. This progression aligns with cognitive development stages and supports more effective acquisition. To support the learning of non-literal meanings, particularly in younger age groups, instructional strategies should include explicit explanations of metaphorical uses, guided practice with contextual cues, and visual or interactive supports to reduce cognitive load. Drawing from approaches in foreign language education (Littlemore, 2004), metaphorical competence in a first-language context can also be fostered through activities that promote analogical reasoning and pattern recognition over time. These strategies may help learners internalize figurative meanings and recognize schematic patterns across similar constructions.

Several limitations of the current study must be acknowledged, alongside recommendations for future research to address these gaps. First, the inclusion of older participants, such as junior high school students, could provide valuable insights into whether the adult-like comprehension and acquisition of the metaphorical meanings of V-diao emerge at later developmental stages. Second, the developmental framework of the GEM could be refined by incorporating a sub-phase within Phase 1 to more precisely account for the linguistic and cognitive abilities of 7-year-old children. Moreover, further comparative analyses of V-diao₃ and V-diao₄ are necessary to uncover their differences, which could shed light on the intricacies of the acquisition process. Finally, alternative classifications of V-diao based on the types of verbs it combines with should be explored. This approach could offer a more comprehensive understanding of its usage and acquisition patterns.