Using a corpus-based multi-dimensional (MD) analysis, this paper investigates the functional linguistic variation in corporate blog (CB), and aims at identifying text types that are distinguished linguistically. Data for a self-built corpus are drawn from 41 out of 50 top ranking corporate blogs. In the multidimensional analysis, 32 linguistic features were tagged, counted and normalized. Then, the normalized frequencies of these features were used in a factor analysis, which identified six factors: (1) personal involvement narration, (2) informational production, (3) interactive persuasion, (4) abstract/impersonal style, (5) evaluative stance, (6) general information vs. technical information. The six factors were then used as predictors in a cluster analysis, through which six clusters were identified, including (1) descriptive discourse, (2) technical conversation, (3) interactive storytelling, (4) evaluative information report, (5) persuasion and (6) interactive instruction. Findings also indicate that all these different clusters (text types) identified have different communicative purposes, but share a common motive for success in business world: using corporate blogs as interactive public relations tools to gain a targeted audience and acquire long-term readers and customers to achieve their organizational goals in business world.
Read full abstract