Cross-Linguistic Analysis of Verb-Noun Preference in Chinese and English: Implications for L2 Chinese Writing

Table of Contents

1. Introduction

Nouns and verbs are fundamental word classes common to all human languages. Research in language acquisition, such as Gentner's (1982) universal noun advantage view, suggests nouns are conceptually easier and acquired earlier. However, cross-linguistic studies reveal significant differences in usage preferences. English exhibits a strong noun preference, especially in formal and academic writing, while Chinese demonstrates a distinct verb preference. This study investigates this contrast empirically using modern newspaper corpora and explores its implications for English-speaking learners of Chinese.

2. Noun/Verb Preference and Ontological Metaphor

The divergence in noun/verb usage is theorized to stem from differing reliance on ontological metaphors (Lakoff & Johnson, 1980). Ontological metaphor involves conceptualizing abstract ideas, emotions, or processes as concrete entities. English frequently nominalizes processes (e.g., "my fear," "her decision"), treating them as manipulable objects. Chinese, in contrast, tends to retain the verbal form to describe states and processes directly (e.g., "I fear," "she decided"). Link (2013) provided preliminary evidence through literary excerpts, but his sample was limited. This study builds upon this theoretical foundation for a systematic, quantitative analysis.

3. Corpus-Based Comparative Study

3.1 Source of Research Materials

To ensure representativeness of modern language use, two corpora were constructed:

• Chinese Corpus: Articles from People's Daily (《人民日报》), a leading official newspaper in China.
• English Corpus: Articles from The New York Times, a major American newspaper.

Articles from the same time period and covering similar topics (e.g., politics, economy, culture) were selected to control for domain variation.

3.2 Research Method and Data Processing

Texts were processed using natural language processing tools for part-of-speech (POS) tagging:

• Chinese: The Stanford CoreNLP Chinese model or Jieba POS tagger was used.
• English: The Stanford CoreNLP English model was used.

Nouns (including common and proper nouns) and verbs (including main verbs and auxiliaries in relevant contexts) were automatically identified and counted. The key metric calculated was the Noun-to-Verb Ratio (NVR):

$NVR = \frac{Count(Nouns)}{Count(Verbs)}$

Statistical tests (e.g., t-tests) were conducted to determine the significance of differences between the corpora.

3.3 Results and Analysis

The analysis confirmed the hypothesized contrast:

4. Impact on English-Speaking Chinese Learners

The study further analyzed writing samples from intermediate-to-advanced English-speaking learners of Chinese. The results showed that these learners' Chinese compositions had an average NVR of approximately 1.8 : 1. This ratio is significantly higher than that of native Chinese writers (closer to 1.1:1) and leans towards the English pattern. This indicates negative transfer from their L1 (English), leading to an underuse of verbs and an over-reliance on nominalized structures in their L2 Chinese writing.

5. Discussion and Pedagogical Implications

The findings have direct implications for Teaching Chinese as a Foreign Language (TCFL):

• Awareness Raising: Instructors should explicitly teach the concept of verb preference in Chinese, contrasting it with English noun preference.
• Input Enhancement: Provide learners with ample authentic materials highlighting natural Chinese verb usage.
• Focused Practice: Design exercises that require converting awkward nominalized phrases (translationese) into more natural verbal constructions.
• Error Correction: Systematically address "nouny" writing in learner feedback.

6. Key Insights

• Empirical Validation: Provides robust, corpus-based evidence for the theoretical verb-noun preference dichotomy between Chinese and English.
• L1 Transfer: Clearly demonstrates how deep-seated L1 grammatical patterns (noun preference) persist in L2 production, affecting stylistic naturalness.
• Beyond Syntax: Highlights that language difference is not merely syntactic but rooted in cognitive styles (ontological metaphor use).
• Pedagogical Gap: Identifies a specific, measurable area (verb usage frequency) often overlooked in traditional grammar-focused instruction.

7. Original Analysis & Expert Commentary

Core Insight: This paper isn't just about counting words; it's a forensic analysis of cognitive style fossilized in grammar. The real story is how English's "noun-centric" worldview, a legacy of its preference for ontological metaphor, creates a persistent stylistic accent in learners of Chinese—an accent that metrics like NVR can now quantify with surgical precision. The study successfully bridges the often-separated worlds of theoretical cognitive linguistics (Lakoff & Johnson) and applied corpus-based SLA research.

Logical Flow: The argument is elegantly linear: Theory (Ontological Metaphor) -> Prior Observation (Link's literary analysis) -> Hypothesis (Modern media will show the same divide) -> Empirical Test (Corpus analysis of NYT vs. People's Daily) -> Confirmation -> Extension (Does L1 transfer affect L2 output?) -> Second Empirical Test (Learner corpus analysis) -> Confirmation -> Practical Implications. This is a textbook example of solid, incremental research design.

Strengths & Flaws: The major strength is its methodological rigor and clear operationalization (NVR). Using comparable newspaper genres controls for register, a common flaw in cross-linguistic studies. However, the analysis has blind spots. First, it treats "noun" and "verb" as monolithic categories. As research from the Universal Dependencies project shows, fine-grained distinctions (e.g., deverbal nouns, light verbs) matter. Does Chinese use more light verb constructions (e.g., 进行讨论) that technically contain a noun but function verbally? This could inflate noun counts. Second, the learner study likely captures ability rather than underlying competence. Do learners over-nominalize because they can't handle complex verbal chains, or is it pure L1 transfer? A think-aloud protocol study could disentangle this.

Actionable Insights: For educators: This study provides the diagnostic tool (NVR) and the treatment plan (contrastive awareness). For technologists: This is a goldmine for AI. Large Language Models (LLMs) like GPT-4 still struggle with generating stylistically native text in a second language. Incorporating a "verb-preference" loss function or fine-tuning on NVR-balanced corpora could significantly improve the naturalness of machine-translated or AI-generated Chinese text, moving beyond mere grammatical correctness. For researchers: The next step is dynamic analysis. Tools like LIWC (Linguistic Inquiry and Word Count) or similar custom dictionaries could track how a learner's NVR evolves over time with targeted instruction, offering a clear metric for pedagogical efficacy.

8. Technical Details & Mathematical Formulation

The core metric, the Noun-to-Verb Ratio (NVR), is a simple but powerful descriptive statistic:

$\text{NVR}_{corpus} = \frac{\sum_{i=1}^{n} N_i}{\sum_{i=1}^{n} V_i}$

Where $N_i$ is the noun count in text sample $i$, and $V_i$ is the verb count in text sample $i$, across $n$ samples in the corpus.

To test for significant differences between two corpora (e.g., Native Chinese vs. Learner Chinese), an independent samples t-test is typically employed:

$t = \frac{\bar{X}_1 - \bar{X}_2}{s_p \sqrt{\frac{1}{n_1} + \frac{1}{n_2}}}$

where $\bar{X}_1$ and $\bar{X}_2$ are the mean NVRs of the two groups, $n_1$ and $n_2$ are sample sizes, and $s_p$ is the pooled standard deviation.

9. Experimental Results & Chart Description

Chart Description (Imagined): A grouped bar chart clearly visualizes the results. The x-axis has three categories: "Native English (NYT)", "Native Chinese (People's Daily)", and "L2 Chinese Learners". The y-axis represents the Average Noun-to-Verb Ratio (NVR).

• The "Native English" bar is the tallest, reaching up to ~2.4.
• The "Native Chinese" bar is the shortest, at ~1.1.
• The "L2 Chinese Learners" bar sits in between, at ~1.8, visually demonstrating the transfer effect—closer to English than to native Chinese.

Error bars (representing standard deviation) on each bar show the variability within each group. Asterisks above the bars indicate statistically significant differences (p < 0.01) between all three groups.

10. Analytical Framework: Case Example

Case: Analyzing a Learner's Sentence

Learner's Output (Translationese): "我对失败的可能性有考虑。" (Literal: "I have consideration for the possibility of failure.")
NVR Analysis: Nouns: 我 (I-pronoun, often counted), 可能性 (possibility), 考虑 (consideration-noun). Verbs: 有 (have). Approx. NVR = 3/1 = 3.0 (Very high, English-like).

Native-like Reformulation (Verb-Preference): "我考虑过可能会失败。" ("I considered that I might fail.")
NVR Analysis: Nouns: 我, 可能 (possibility?). Verbs: 考虑过 (considered), 会 (might), 失败 (fail). Approx. NVR = 2/3 ≈ 0.67 (Low, verb-heavy).

This micro-case shows how the analytical framework pinpoints the exact location of L1 interference—the nominalization of "考虑" (consideration) and the use of a possessive structure—and guides its correction towards a more natural verbal construction.

11. Future Applications & Research Directions

• AI & NLP: Integrate NVR and similar stylistic metrics into evaluation benchmarks for machine translation and text generation. Develop style-transfer models specifically trained to adjust the "nouniness" of output text to match target language norms.
• Adaptive Learning Platforms: Create writing assistants that provide real-time feedback on stylistic metrics like NVR, helping learners gradually shift their output towards target-language norms.
• Neurolinguistics: Use fMRI or EEG to investigate if processing high-NVR (nouny) Chinese sentences activates different brain regions in L2 learners compared to native speakers, linking behavioral patterns to neural processing.
• Broader Cross-Linguistic Studies: Apply this framework to other language pairs (e.g., German vs. Spanish, Japanese vs. Korean) to map a typology of "noun-biased" vs. "verb-biased" languages and refine the ontological metaphor theory.
• Longitudinal Studies: Track learners over years to see if NVR naturally converges with native norms through immersion or if explicit instruction is necessary for lasting change.

12. References

1. Biber, D., Conrad, S., & Reppen, R. (1998). Corpus linguistics: Investigating language structure and use. Cambridge University Press.
2. Gentner, D. (1982). Why nouns are learned before verbs: Linguistic relativity versus natural partitioning. In S. A. Kuczaj II (Ed.), Language development: Vol. 2. Language, thought, and culture (pp. 301–334). Erlbaum.
3. Lakoff, G., & Johnson, M. (1980). Metaphors we live by. University of Chicago Press.
4. Link, P. (2013). An anatomy of Chinese: Rhythm, metaphor, politics. Harvard University Press.
5. Tardif, T. (1996). Nouns are not always learned before verbs: Evidence from Mandarin speakers' early vocabularies. Developmental Psychology, 32(3), 492–504.
6. Tardif, T., Gelman, S. A., & Xu, F. (1999). Putting the "noun bias" in context: A comparison of English and Mandarin. Child Development, 70(3), 620–635.
7. Zhu, Y., Yan, S., & Li, S. (2021). International Journal of Chinese Language Teaching, 2(2), 32-43. (The analyzed paper).
8. Universal Dependencies Consortium. (2023). Universal Dependencies. https://universaldependencies.org/
9. Pennebaker, J.W., Boyd, R.L., Jordan, K., & Blackburn, K. (2015). The development and psychometric properties of LIWC2015. University of Texas at Austin.