This monograph is Paul Baker's most recent book-length attempt to demonstrate how corpora could be used to address sociolinguistic and societal issues. Systematically illustrating corpus-enabled analytical methods throughout the book, he makes an excellent case for the application of corpora in researching gendered language and gendered discourse.
After problematizing the ‘difference paradigm’ in the introduction, chapter 2 reports on a corpus-driven study examining male and female keywords in the spoken part of the British National Corpus and concludes that gendered language differences are sometimes overstated. Chapter 3 is a corpus-based study focusing on the disagreement strategies employed by female speakers in the Michigan Corpus of Academic Spoken English. Although four strategies (‘bald-online’, ‘downtoner’, ‘indirectness’, and ‘positive politeness’) are identified, no consistent pattern could be generalized among all female speakers. Baker, therefore, cautions against the concept of ‘female style’ in speech and maintains that other factors (such as the relationship between interlocutors) matter as well as gender.
Turning to a diachronic perspective on gendered discourse, chapter 4 investigates the representation of males and females in the Corpus of Historical American English. Baker discusses the implication of the frequency gap between MAN and WOMAN, gender-marked social roles, familial terms, ‘male firstness’ in binominal pairs, and the generic use of man. They all point to a male bias in American society, although the increasing frequency of gender-inclusive terms might suggest a more gender-balanced mindset today. Chapter 5 compares the representation of gay men in the Daily Mail at two time points. Examining the collocates and the discourse prosody of gay and homosexual, Baker finds that negative discourses are decreasing and also becoming more nuanced, and that gay rights are exploited by the tabloid to tell ‘its readers whom to dislike’ (131). The author also discusses the legitimization strategies employed by the tabloid to voice negativity towards homosexuality. In chapter 6, Baker uses Sketch Engine to find the statistically significant collocates of BOYS and GIRLS and concludes that the gendered representations reflected by the collocates are consistent with gendered stereotypes in society.
Chapter 7 demonstrates how multiple corpus techniques and multiple corpora could be used to triangulate research findings. The author compiles three corpora from the website Craigslist where heterosexual men posted adverts to seek women in Australia, India, and Singapore. The semantic tags in the corpora and the collocational networks of self-descriptors show that men in these countries represented themselves in fairly similar ways, although slightly different emphases are also noticeable. The analysis of the concordance lines containing WOMEN reveals problematic discourses constructed against women. In the concluding chapter, Baker provides a reflexive account to remind readers of the limitations of the corpus approach and the importance of triangulating corpus findings with contextualized knowledge. He also calls upon researchers' efforts to engage a larger audience and to disseminate findings in gender and language to those outside the academic circle.
