publisher colophon

Related Article: https://muse.jhu.edu/article/757632

The figures from the main article are repeated here in color, with their captions.

Figure 1 (p. 389). Dependency length as a function of sentence length for fifty-three languages. The x-axis is sentence length and the y-axis is the mean of total dependency length for all sentences of that length. The black line represents true dependency length. The colored lines represent random baselines. Solid colored lines represent projective baselines; dashed colored lines represent nonprojective baselines. Some baselines are not visible because they are covered by others.
Click for larger view
View full resolution
Figure 1 (p. 389).

Dependency length as a function of sentence length for fifty-three languages. The x-axis is sentence length and the y-axis is the mean of total dependency length for all sentences of that length. The black line represents true dependency length. The colored lines represent random baselines. Solid colored lines represent projective baselines; dashed colored lines represent nonprojective baselines. Some baselines are not visible because they are covered by others.

Figure 2 (p. 390). Histograms of languages by proportion of sentences with shorter dependencies than average baseline reorderings. One histogram per baseline. The vertical line indicates 50% on the x-axis. Darker gray blocks indicate languages where the significant majority of sentences have shorter dependencies than the baseline. Light gray indicates languages where the proportion of such sentences is not significantly different from one half. Black indicates languages where the significant majority of sentences in the corpus have longer dependency length than the baseline. The black exceptions are corpora of Uyghur (proportion of optimized sentences = 0.46), Latin (0.45), and Northern Kurdish (0.39) when compared to the fixed random baseline, and Korean when compared to the fixed (0.46) and free (0.47) head-consistent random baselines.
Click for larger view
View full resolution
Figure 2 (p. 390).

Histograms of languages by proportion of sentences with shorter dependencies than average baseline reorderings. One histogram per baseline. The vertical line indicates 50% on the x-axis. Darker gray blocks indicate languages where the significant majority of sentences have shorter dependencies than the baseline. Light gray indicates languages where the proportion of such sentences is not significantly different from one half. Black indicates languages where the significant majority of sentences in the corpus have longer dependency length than the baseline. The black exceptions are corpora of Uyghur (proportion of optimized sentences = 0.46), Latin (0.45), and Northern Kurdish (0.39) when compared to the fixed random baseline, and Korean when compared to the fixed (0.46) and free (0.47) head-consistent random baselines.

Figure 3 (p. 392). Schematic for how grammar and usage relate to linearizations. Grammar selects a set of permitted linearizations from the logically possible ones; usage selects one linearization from the grammatically permitted ones.
Click for larger view
View full resolution
Figure 3 (p. 392).

Schematic for how grammar and usage relate to linearizations. Grammar selects a set of permitted linearizations from the logically possible ones; usage selects one linearization from the grammatically permitted ones.

Figure 4 (p. 395). Mean dependency length as a function of sentence length, for real linearizations as compared with grammatical baselines.
Click for larger view
View full resolution
Figure 4 (p. 395).

Mean dependency length as a function of sentence length, for real linearizations as compared with grammatical baselines.

Figure 5 (p. 397). Languages by proportion of head-final dependencies (x-axis) and mean dependency length per word at sentence lengths 10, 15, and 20 (y-axis). Languages are represented by their two- or three-letter ISO codes. Pearson (r) and Spearman (rho) correlation coefficients are given.
Click for larger view
View full resolution
Figure 5 (p. 397).

Languages by proportion of head-final dependencies (x-axis) and mean dependency length per word at sentence lengths 10, 15, and 20 (y-axis). Languages are represented by their two- or three-letter ISO codes. Pearson (r) and Spearman (rho) correlation coefficients are given.

Figure 6 (p. 398). Languages by proportion of head-final dependencies (x-axis) and mean dependency length per word at sentence lengths 10, 15, and 20 (y-axis), . Pearson and Spearman correlation coefficients are given.
Click for larger view
View full resolution
Figure 6 (p. 398).

Languages by proportion of head-final dependencies (x-axis) and mean dependency length per word at sentence lengths 10, 15, and 20 (y-axis), FOR CONTENT WORDS ONLY. Pearson and Spearman correlation coefficients are given.

Figure A1 (p. 402). Mean dependency length as a function of sentence length for fifty-three languages, . The black line represents true dependency length. The colored lines represent random baselines. Solid colored lines represent projective baselines; dashed colored lines represent nonprojective baselines. Some baselines are not visible because they are covered by others.
Click for larger view
View full resolution
Figure A1 (p. 402).

Mean dependency length as a function of sentence length for fifty-three languages, ACCORDING TO ORIGINAL UD PARSES. The black line represents true dependency length. The colored lines represent random baselines. Solid colored lines represent projective baselines; dashed colored lines represent nonprojective baselines. Some baselines are not visible because they are covered by others.

Figure A2 (p. 403). Histograms of languages by proportion of sentences with shorter dependencies than average baseline reorderings, . Shading means the same as in Fig. 2. The black exceptions are corpora of Ancient Greek (proportion of optimized sentences = 0.49) and Latin (0.45) when compared to the fixed random baseline.
Click for larger view
View full resolution
Figure A2 (p. 403).

Histograms of languages by proportion of sentences with shorter dependencies than average baseline reorderings, ACCORDING TO ORIGINAL UD PARSES. Shading means the same as in Fig. 2. The black exceptions are corpora of Ancient Greek (proportion of optimized sentences = 0.49) and Latin (0.45) when compared to the fixed random baseline.

Figure A3 (p. 404). Mean dependency length as a function of sentence length for fifty-three languages, . The black line represents true dependency length. The colored lines represent random baselines. Solid colored lines represent projective baselines; dashed colored lines represent nonprojective baselines. Some baselines are not visible because they are covered by others.
Click for larger view
View full resolution
Figure A3 (p. 404).

Mean dependency length as a function of sentence length for fifty-three languages, IGNORING ALL FUNCTION WORDS. The black line represents true dependency length. The colored lines represent random baselines. Solid colored lines represent projective baselines; dashed colored lines represent nonprojective baselines. Some baselines are not visible because they are covered by others.

Richard Futrell
University of California, Irvine
Roger P. Levy
Massachusetts Institute of Technology
Edward Gibson
Massachusetts Institute of Technology

Additional Information

ISSN
1535-0665
Print ISSN
0097-8507
Launched on MUSE
2020-06-19
Open Access
Yes
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.