Wanna-Contraction Acceptability

Introduction

Wanna (want to) is an English contraction whose use depends on specific syntactic restrictions. In some wh-questions, contraction is permitted, as in who do you wanna meet? . However, in other cases it is no acceptable, as in who do you want to meet to? . This difference is explained by a well-known syntactic constraint on wanna-contraction.

Contraction is not possible in wh-questions when the wh-word functions as the subject of the infinitival clause. In these cases, want to can’t be reduced to wanna. When the wh-word is not the subject of the infinitival clause, contraction is allowed. Therefore, the availability of wanna-contraction depends on whether a subject intervenes between want and to.

This analyses examines data from four participants groups: L1 adults, L1 children, adult L2 learners, and child L2 learners. Each participant judged sentences that differed in clause type (if vs. who) and gap condition (gap vs. not gap). A gap refers to an unfilled grammatical position in the sentence, which may be a subject or an objec, while no gap indicates that all positions are filled.

1. Acceptability Across Groups

Mean acceptability judgments were plotted for every combination of clause type (if Vs. Who) and condition (gap Vs. no gap) to compare how participant responded to different sentences types. In the following graph, the bars show the average proportion of acceptable responses per group. The figure indicates that the four groups responded differently to the conditions.

library(tidyverse)
library(ggplot2)
library(readxl)


OSF <- OSF %>%
 arrange(group)

dat_summary <- OSF %>%
group_by(group, clause, gap) %>%
summarise(mean_accept = mean(judgment, na.rm = TRUE))

dat_summary$group <- factor(dat_summary$group,
                            levels = c("L1_adults", "L2_adults", 
                                       "L1_children", "L2_children"))

ggplot(dat_summary, aes(x = gap, y = mean_accept, fill = clause)) +
  geom_col(position = "dodge") +
  facet_wrap(~ group, nrow=2) +
  labs(title = "Acceptability Across Groups", x = "Gap Condition", y = "Mean Acceptability"
  )

Overall, this column graph displays that if no gap sentences received higher acceptability rating than the the gap sentence specifically across adult groups, who showed low ratings for if gap sentences. In contrast who gap sentences were rated higher than if gap sentences.

2. Gap-only: if vs. who (points + error bars)

Are who-sentences more accepted than if-sentences when there is a gap?

The table below reports mean acceptability ratings for if and who clauses in the gap condition across participant groups.

  gap_data <- OSF %>%
  filter(gap == "gap")
gap_data %>%
  group_by(group, clause) %>%
  summarise(
    mean_accept = mean(judgment, na.rm = TRUE),
    .groups = "drop"
  )

gap_data %>%
  group_by(group, clause) %>%
  summarise(
    mean_accept = mean(judgment, na.rm = TRUE),
    se_accept   = sd(judgment, na.rm = TRUE) / sqrt(n()),
    .groups = "drop"
  ) %>%
  ggplot(aes(x = clause, y = mean_accept)) +
  geom_point(size = 3) +
  geom_errorbar(
    aes(
      ymin = mean_accept - se_accept,
      ymax = mean_accept + se_accept
    ),
    width = 0.15
  ) +
  facet_wrap(~ group) +
  labs(
    x = "Clause Type",
    y = "Mean Acceptability (gap condition)"
  ) +
  theme_minimal()

This graph visualizes these means on the table and includes standard errors that facilitates comparison across groups.

Across groups, who-clauses receive higher acceptability ratings than if-clauses in the gap condition. This contrast is strongest in L1 adults who rated who-gap sentences very high and if-gap sentences very low. L2 adults and children show the same pattern, but the difference is smaller, suggesting weaker sensitivity to the syntactic constraint on wanna contraction.

Conclusion

This analyses shows a consistent preference for who-clauses over if-clauses. This effect is strongest among L1 adults, while the effect is weaker in L2 learners and children.