List of figures

Part of

Textbook English: A multi-dimensional approach
Elen Le Foll
[Studies in Corpus Linguistics 116] 2024
► pp. xiii–xvi

List of figures

Figure 1.Cumulative word counts of the Conversation, Fiction, Instructional, Informative, and Personal correspondence subcorpora of the TEC per textbook volume85

Figure 2.Correlation matrix of the normalised counts in Table 998

Figure 3.Mean scores of general spoken and written registers of English on Biber’s (1988) Dimension 1 (as summarised in Biber & Conrad 2019: 292)107

Figure 4.Comparison of the conversation, fiction and informative texts from the TEC with the three corresponding target language reference corpora on Biber’s (1988) Dimension 1 (as calculated by the MAT) (Le Foll 2021. Zenodo.

)109

Figure 5.Per-feature accuracy measures (and bootstrapped 95% confidence intervals) of the MFTE on samples of the TEC, the Spoken BNC2014 and data comparable to the Youth Fiction and Info Teens (for details, see Le Foll 2022c: 277–81)121

Figure 6.Distribution of normalised frequencies of five features across the TEC (histograms) and visualisations of their correlations (scatterplots)128

Figure 7.Dimension 1 mean scores for disciplines (left) and genre families (right) from Gardner et al. (2019: 655)137

Figure 8.Scree plot of the eigenvalues of the principal components (PCs) for the TEC data141

Figure 9.Snapshots from the 3-D visualisation of the first three dimensions of the multi-dimensional model of intra-textbook variation142

Figure 10.Scatterplot matrix of all the combinations of the first six dimensions of the model of intra-textbook variation143

Figure 11.Projection of the texts of the TEC on the first and second dimensions of the model of intra-textbook variation144

Figure 12.Graph of the features with the strongest contributions to the first and second dimensions of the model of intra-textbook variation (see also Table 17)145

Figure 13.Projection of the texts of the TEC on the third and fourth dimensions of the model of intra-textbook variation147

Figure 14.Graph of the features with the strongest contributions to the third and fourth dimensions of the model of intra-textbook variation148

Figure 15.Projection of the texts of TEC on the third and fourth dimensions with colours and ellipses indicating the proficiency level of the textbooks (as opposed to register as in Figure 13)149

Figure 16.Coefficient estimates and 95% confidence intervals of the fixed effects in the model: lmer(PC1 ~ Register + (1|Series))154

Figure 17.Coefficient estimates and 95% confidence intervals of the fixed effects in the model: lmer(PC2 ~ Register + Level + Register*Level + (1|Series))157

Figure 18.Estimated PC2 scores across each register and the five textbook proficiency levels159

Figure 19.Coefficient estimates and 95% confidence intervals of the fixed effects in the model: lmer(PC3 ~ Register + Level + Register*Level + (1|Series)162

Figure 20.Estimated PC3 scores across each register and the five textbook proficiency levels163

Figure 21.Coefficient estimates and 95% confidence intervals of the fixed effects in the model: lmer(PC4 ~ Register + Level + Register*Level + (1|Series))164

Figure 22.Estimated PC4 scores across each register and the five textbook proficiency levels165

Figure 23.Scree plot of the eigenvalues of the PCs for the Textbook English vs. ‘real-world’ English PCA168

Figure 24.Snapshots from the 3-D representation of texts along PC1–PC3169

Figure 25.Scatterplot matrix of combinations of the four dimensions of the model of Textbook English vs. ‘real-world’ English170

Figure 26.Projection of the texts of the three subcorpora of the TEC and the reference corpora on PC1 and PC2171

Figure 27.Graph of the features with the strongest contributions to the first and second dimensions172

Figure 28.Projection of the texts of the three subcorpora of the TEC and the reference corpora on PC3 and PC4175

Figure 29.Projection of the texts of the three subcorpora of the TEC and the reference corpora on PC3 and PC4 with ellipses representing the five textbook proficiency levels vs. the reference corpora176

Figure 30.Predicted PC3 scores of the texts of the TEC and the reference corpora178

Figure 31.Graph of the features with the strongest contributions to the third and fourth dimensions179

Figure 32.Predicted PC4 scores of the texts of the TEC and the reference corpora181

Figure 33.Predicted PC1 scores of the texts of the TEC and the reference corpora182

Figure 34.Predicted PC2 scores of the texts of the TEC and the reference corpora191

Figure 35.Normalised counts of selected features with salient loadings on PC1 in the Textbook Informative subcorpus (Levels A to E) and the reference Info Teens corpus (Ref.)192

Figure 36.Word frequency analysis conducted with english-corpora.org (on the basis of COCA data) of Excerpt (59)220

Figure 37.Part of the ‘word profile’ page of the word lush as generated on english-corpora.org/coca221

Figure 38.Word frequency analysis conducted with english-corpora.org (on the basis of the COCA) of Excerpt (56)221

Figure 39.Part of the “word profile” page of the word moreover on english-corpora.org/coca222

Figure 40.Projection of texts on PC1 and PC2 from a random 2/3 split-data analysis of the three subcorpora of the TEC and the three reference corpora239

Figure 41.Projection of texts on PC3 and PC4 from a random 2/3 split-data analysis of the three subcorpora of the TEC and the three reference corpora240