Previously, and by that we mean about 6 months ago, we covered the various methods where one could use to assess linguistic diversity, which are extensions of biodiversity evaluation methods contextualised in linguistics. We have covered anything from richness to the Greenberg index, and looked at the downsides of evaluating linguistic diversity using these indices.

But there is also another approach to assessing biodiversity, one that takes into account the spatial patterns of species richness in different habitats. Robert H. Whittaker would propose the three levels of species diversity — alpha, beta, and gamma. While first mentioned in the 1960 paper titled Vegetation of the Siskiyou Mountains, Oregon and California, this would lead to the 1972 paper called Evolution and Measurement of Species Diversity, which dives into detail each individual level of species diversity.

Briefly put, alpha diversity is the mean of the observed species diversity within a certain set of predefined sites. In other words, alpha diversity measures the mean local diversity within a set number of given areas assessed. It translates somewhat nicely into linguistic diversity — imagine you want to assess linguistic diversity in specific cities like Madrid, Barcelona, Bilbao, Murcia, Vigo, and Zaragoza. The alpha diversity of these selected cities would be the mean of the linguistic diversity reported in these areas.

How do we measure this local diversity, you may ask. While more crude analyses might tend to use any of the diversity indices mentioned previously, Grin and Fürst proposed that the Shannon index be used. A weighted mean would be taken for all the sites considered, with areas with greater population sizes given more weight.

Beta diversity takes it a step further. It compares the diversity between a certain number of predefined sites. The more varied these sites are, the more diversity this system has. So, if these specific cities in Spain have a high beta diversity, that would indicate that these Spanish cities are more diverse as a whole.

Gamma diversity gives us diversity of the big picture. Going back to the Spanish cities example, the gamma diversity of these selected areas could give us an idea of the diversity of Spain’s languages. Calculating gamma diversity takes on two predominant forms — the more common one takes a product (or multiplication) of the alpha and beta diversities, while the most common alternative approach uses a sum of the alpha and beta diversities.

According to Grin and Fürst, a multi-level approach to assessing linguistic diversity has important implications on language policies in a certain country or region. For example, if a certain policy to increase language diversity in every city in a country is implemented, we could see the alpha diversity of languages increase, as we would expect from this policy. However, we could also see a decrease in beta diversity as this policy is implemented countrywide. In other words, between different cities, we could see the diversities becoming more homogenous, despite having more languages spoken within these cities. In order words, a multi-level approach allows a more in-depth look in diversity structures in a particular linguistic region.

Another advantage of such a multi-level approach is the interpretations that can be drawn from a more spatiotemporal angle. Immigration, emigration, and demographic change can all result in changes in linguistic diversity. However, if we use a single-level approach, we might not be able to detect the finer changes in diversity structures, for example, in a within- or between-neighbourhood scale. This way, we could see if certain urban policies could result in a more linguistically segregated situation between neighbourhoods, marked by a decrease in alpha diversity, but an increased beta or gamma diversity, for example.

One of the questions I had in mind from the previous essay was how multilingualism would be accounted for. It is possible for individuals to be native speakers or L1 speakers of 2 or more languages, which can affect diversity indices as pointed out previously. As the authors noted in their discussion, these assessments hinge on the assumption of monolingualism, and that if there are multilingual individuals, they would tend to choose one of their native languages to satisfy the assumption of monolingualism. One suggestion was to include bilingualism and so on as individual linguistic repertoires, which could result in a higher gamma diversity as there are more combinations of linguistic repertoires than there are possibly languages spoken in an assessed area.

Overall, it seems that the primary benefit of a multi-level approach of evaluating linguistic diversity is the development of a more efficient system of resource allocation when it comes to managing linguistic diversity or language education. This is paralleled by the resource allocation to conserving or managing biodiversity in certain regions. Depending on the scope of language policies, such alpha diversities could be assessed using scales from as small as that of the neighbourhood or small village, to an administrative division like a county.

When taken further, this could inform policies on education of indigenous languages, a topic of particular concern due to the stark proportion of languages at risk of extinction. Similarly, language revitalisation efforts in certain areas could also be informed, to preserve the linguistic diversity in more disproportionately affected regions. It would be interesting to look into further developments and extensions made using this approach, as these diversities are contextualised in languages and linguistics, and to observe its real-world applications and informing policymaking.

Further Reading

Grin, F., Fürst, G. Measuring Linguistic Diversity: A Multi-level Metric. Soc Indic Res 164, 601–621 (2022). https://doi.org/10.1007/s11205-022-02934-5.