The body of case study-based research on language in multinational corporations (MNCs) is growing, but its findings have as yet been subjected to limited statistical validation. In this paper we use quantitative functional-level data to chart language use in subsidiaries’ communication with other MNC units and local partner firms, and to analyze some consequences of these patterns against the background of previous qualitative work in the area. Our findings confirm that MNCs are indeed multilingual, but that language fluency varies significantly across functions and organizational levels. This has important implications for communication, knowledge sharing and the viability of formal language strategies.
It is by now widely accepted that language in multinational corporations (MNCs) merits study as a stand-alone topic rather than simply a component of cultural distance, and that the vast majority of multinationals are in fact multilingual (Barner-Rasmussen and Björkman, 2005, Luo and Shenkar, 2006 and Piekkari, 2006). These developments are welcome because they help us understand some key day-to-day realities of MNC management and help bring MNC research closer to reality. Specifically, these findings mark a shift away from treating the MNC as a ‘monolithic’ and unified, raising instead important questions related to internal fragmentation (e.g., Kristensen and Zeitlin, 2001, Kristensen and Zeitlin, 2005 and Morgan and Kristensen, 2006), the limits of top management agency, and the central planning ability of corporate centres.
We nevertheless argue that the full implications of these new insights have yet to be fully digested. Specifically, we question whether top management can meaningfully ‘design’ a ‘language system’ for the MNC and ‘align’ it with organizational strategy, as suggested by Luo and Shenkar (2006). Such ideas can be seen as typical of what Janssens, Lambert and Steyaert (2004) term the ‘mechanistic’ perspective on language in international management, with the limitations and potential pitfalls that this perspective entails. It can be argued that taking a mechanistic perspective on MNC management is unrealistic, and as such unproductive – perhaps even dangerous. Again there are parallels to the discussion on the limits of top management agency, as voiced by e.g., Kristensen and Zeitlin, 2001 and Kristensen and Zeitlin, 2005.
These dangers of language research in the MNC context may be accentuated by the fact that the bulk of empirical work in the area thus far is case study-based (Tietze, 2007). Case studies are commonly considered useful to grasp emerging research areas and generate hypotheses that can subsequently be tested by other means (see e.g., Eisenhardt, 1989). They have undoubtedly yielded extremely valuable insights into MNC language dynamics. It is nevertheless difficult to refute the argument that large-scale quantitative studies would at this point provide useful descriptive information that has not been available before, and confer empirical stability upon the diverse claims that are being made.
Against this background, our aim with this paper is to provide an empirical exploration of language diversity and the variation in language skills inside multinationals. We also contrast our results with previous work in the area (particularly the theoretical propositions presented by Luo & Shenkar, 2006) and discuss the outcome in terms of how MNCs can realistically be understood and managed as multilingual organizations. We fulfill this aim by conducting a quantitative hypothesis-testing study based on survey data from 61 Finnish subsidiaries of foreign MNCs. By focusing on the functional level of analysis, we provide a more fine-grained study than any previous quantitative piece.
