"For evangelization purposes, a people group is the largest group within which the Gospel can spread as a church planting movement without encountering barriers of understanding or acceptance" (Source: 1982 Lausanne Committee Chicago meeting.)
In most parts of the world lack of understandability acts as the main barrier and it is appropriate to define people group primarily by language, with the possibility of sub-divisions based on dialect or cultural variations. Such a list may be referred to as an ethno-linguistic list of peoples.
"For evangelization purposes, a people group is the largest group within which the Gospel can spread as a church planting movement without encountering barriers of understanding or acceptance"
In other parts of the world, most notably in portions of South Asia, acceptance is a greater barrier than understandability. In these regions, caste, religious tradition, location, and common histories and legends may be used to identify the primary boundary of each people group. Language can be a secondary boundary.
Ideally, "people group" would always be defined to mean (a) all individuals in the group understand each other reasonably well and (b) cultural / relationship barriers aren't so high that the transmission of the Gospel is seriously impeded.
This is the goal, and this is what Joshua Project strives towards. However, there are situations where compromises are in order.
Let's consider the most complicated case, India. If we strictly defined "people group" this way in India, all individuals in each people group would speak the same language and would be in the same caste or tribe. Easily done, but what would it lead to? We would immediately create over 16,000 people groups in India alone, using the excellent Omid data for India. (Some castes speak 50-60 languages.) This would be overwhelming, from a ministry perspective. Is this the best way to serve church planters and other workers on the field?
An argument can be made for defining "people group" according to the highest of the two barriers (understandability and acceptance), in some circumstances. If understandability is the most important barrier, then a linguistic, or an ethno-linguistic, approach is used. With this approach, one people group doesn't speak more than one language (apart from occasional bi-lingual individuals), although more than one people group may speak a given language if cultural or dialect differences warrant. In most of the world, this is the approach Joshua Project uses.
If the cultural / relationship barrier is the greatest barrier (as it often is in South Asia), then we treat caste / tribe as the first criterion for separation. With this approach, one people group may speak more than one language. And as with the first approach, one language may be spoken by more than one people group. (Note that multi-lingualism is not so much in view as the situation where some individuals in the group speak language A and others speak language B. Some may be multi-lingual, but that's not the main issue.)
Allowing one people group to speak more than one language is a compromise. It is a way to present a somewhat simplified picture of the church planting task, at the risk of over-simplifying the understandability barriers within people groups.
Joshua Project's first purpose is to support church planting and discipleship efforts, and we define "people group" with that goal in mind.
The likely alternative is to always present a list of language groups, with some sub-divisions of language groups (an ethno-linguistic list). The first cut is always on language, using this alternative. Consequently, church planters and disciplers focus on ethno-linguistic groups, attempting to plant churches (oftentimes) among castes and tribes that mistrust or dislike each other. People may understand each other, but do they accept each other?
Joshua Project's first purpose is to support church planting and discipleship efforts, and we define "people group" with that goal in mind. We also wish to support language-based ministry efforts (such as literature / video / audio distributions and the like), as a more secondary purpose. Our data can be used for both purposes. We can format the data to support church planting, and to some extent we can re-format it to support language-based outreach.
How a people group is defined will determine how many people groups there are. Different definitions yield different counts. For more details see How Many People Groups are There?.