Verb-Second Effect on Quantifier Scope Interpretation

Asad Sayeed, Matthias Lindemann, Vera Demberg2019 · DOI

TL;DR

This paper investigates how word order and world knowledge interact to resolve ambiguity in quantifier scope, a common linguistic phenomenon. By leveraging the flexible verb-second word order of German, the researchers designed a web-based experiment that suggests syntactic structure, particularly word order variation, plays a more significant role than previously thought, especially when dealing with indefinite quantifiers.

What Problem Does This Paper Address?

Many sentences can have multiple interpretations due to the order of quantifiers. For example, "Every child climbed a tree" can mean each child climbed their own tree (linear scope) or there's one specific tree that all children climbed (inverse scope). While linguists understand how these ambiguities are formally possible, how humans resolve them in real-time is less clear. Previous research has explored the roles of semantic reanalysis and background world knowledge. However, studies often use English, where word order is relatively fixed, making it hard to isolate the effect of word order itself. This paper aims to disentangle the influence of word order variation from world knowledge in resolving quantifier scope ambiguities, particularly focusing on how languages with more flexible word order, like German, might shed light on this.

Methods

The researchers designed a web-based cloze (fill-in-the-blank) experiment in German. They exploited German's verb-second (V2) word order, which allows for more flexibility in placing sentence elements compared to English. Specifically, they created sentences where the object (containing a quantifier) could precede the subject (also containing a quantifier), creating an Object-Subject-Verb (OSV) order, as opposed to the more common Subject-Verb-Object (SVO) order. Participants were presented with a "context" sentence and then a "continuation" sentence with a blank for the subject and verb. The task was to fill in the blank, and their choices were analyzed to determine whether they interpreted the original sentence with a singular or plural scope. This "context-continuation" paradigm is considered more sensitive to online processing than asking for explicit interpretations.

Datasets

The study used stimuli translated from English experiments (specifically Dwivedi, 2013) and adapted for German. The experiment was hosted on a platform called Lingoturk, which is designed for developing and hosting web-based psycholinguistic experiments. Participants were recruited from undergraduate students at a German-speaking university and via the online platform Prolific. The collected data consisted of 1546 responses, with 1236 being unambiguous and interpretable references to the object of the context sentence.

Key Insights and Ideas

  • German's V2 word order is a powerful tool: The flexibility of German's verb-second syntax allows for systematic manipulation of word order (SVO vs. OSV) in a way that is not easily achievable in English, providing a cleaner way to study the effect of linear order.
  • Word order matters, especially with indefinites: While definite articles (singular or plural) strongly dictated the interpretation regardless of word order, indefinite articles showed a significant interaction with word order. This suggests that the processor is more likely to adjust its interpretation based on word order when dealing with indefinite quantifiers.
  • Challenging the "world knowledge first" hypothesis: The results suggest that syntactic structure, particularly word order, can exert a stronger influence on scope interpretation than previously thought, especially in languages where word order variation is common. This pushes back against the idea that background world knowledge always dominates interpretation.
  • German speakers are more attuned to word order effects: The authors hypothesize that because German speakers are more accustomed to encountering varied word orders, they are more likely to notice and be influenced by syntactic cues like linear order, making them less reliant on purely pragmatic or world-knowledge-based interpretations.

Key Results

A statistical model (binomial logit regression) revealed a significant effect of word order on plural interpretation for sentences with indefinite articles. Specifically, the SVO order was associated with a higher probability of a plural interpretation compared to the OSV order (log-odds ratio b = 0.926). This indicates that when the indefinite object appears first (OSV), it is less likely to be interpreted as plural. Definite articles, however, showed little to no effect of word order on interpretation.

Limitations and Future Work

The authors acknowledge that the cloze task, while offering ecological validity, resulted in about 20% of the data being discarded due to uninterpretable or ambiguous completions. They also note that their experiment focused on a specific type of ambiguity and language. Future work includes replicating these findings using other experimental modalities like self-paced reading and eye-tracking to capture the temporal dynamics of scope resolution. They also plan to use corpus analysis and machine learning to further investigate the empirical basis of plurality judgments in ambiguously scoped events.

Summary updated

Abstract

Sentences like "Every child climbed a tree" have at least two interpretations depending on the precedence order of the universal quantifier and the indefinite. Previous experimental work explores the role that different mechanisms such as semantic reanalysis and world knowledge may have in enabling each interpretation. This paper discusses a web-based task that uses the verb-second characteristic of German main clauses to estimate the influence of word order variation over world knowledge.
BibTeX
@misc{sayeed2019verb,
  author = {Asad Sayeed and Matthias Lindemann and Vera Demberg},
  title = {{Verb-Second Effect on Quantifier Scope Interpretation}},
  year = {2019},
  doi = {10.18653/v1/w19-2915},
}
Figure