Gilbert’s Contribution to Group Ontology

Margaret Gilbert’s “Walking Together: A Paradigmatic Social Phenomenon” is a seminal paper in the analytic debate on joint action. On 14 pages it summarises what Gilbert considered the most important insights from her book On Social Facts.

The paper has earned its position in the literature. It was a path-breaking contribution and remains worth reading more than two decades after publication. But to all students out there: Do not start your papers like that! First Gilbert presents her paper as a contribution to the philosophy of the social sciences, then she frames it as an analysis of groups, before she turns to a rather special case of joint action. In other words, Gilbert first asks a question about the social sciences, then switches to a question about social groups in general, to then actually discuss a special case of social groups: The group of two people walking together. Continue reading “Gilbert’s Contribution to Group Ontology”


On Organised and Feature Groups

For my upcoming talk on group ontology I am re-reading key papers on the topic. One of the most recent contributions is Brian Epstein’s “What Are Social Groups?“. Brian wrote one of the most advanced and wide-ranging text on the topic, but I will focus on a minor point from his paper.

Right at the beginning Brian discusses Katherine Ritchie’s distinction between organised and features groups. Microsoft is an organised group and males are a feature group. An organised group is characterised by a structure and a feature group by a feature, such as being male. You are a member of Microsoft if you fill a node in Microsoft’s structure and you are a member of the group males if you have the property to be male. Ritchie complicates the analysis slightly by arguing that the feature must be socially constructed, but that will remain secondary for my post.

Brian raises a number of serious problems for Ritchie’s distinction between the two group types. I look at one of those problems and suggest that it is not a problem after all. Here is the central quote:

“A key challenge for this approach is how to understand a “feature” in the latter category. Which sorts of features that members possess count for such groups, and which are ruled out? Ritchie needs to balance this carefully: if we include all properties, including extrinsic ones, then even the property being a person filling in a node of such-and-such a structure counts, so all groups would be feature groups and the intended contrast between the categories would collapse. If, on the other hand, the “features” were restricted to only intrinsic properties, then we would leave out the archetypal groups Ritchie highlights, such as races and genders.” (p. 4)

Brian argues that filling a node in a certain structure is a property and that therefore you cannot distinguish organised groups from feature groups.

The first few times I read the quoted passage Brian convinced me. Since then I’ve changed my mind. I now look at it this way: Brian is right, to fill a node of such-and-such a structure is a feature. I even grant it is the kind of feature that individuates a feature group. So there is a feature group of those individuals who have the feature to fill nodes of such-and-such a structure. There is the feature group of people who fill the nodes of Microsoft’s corporate structure, which I call the Microsoft-feature-group for short.

But admitting this feature group does not undermine the distinction between feature and structure groups at all! There are just two groups: an organised group and a feature group. There is the group Microsoft, an organised group, and there is the Microsoft-feature-group. They are two groups of two types.

Sure, being a member of Microsoft entails being a member of the Microsoft-feature-group and vice versa. But why is that a problem? The two groups share all their members at all times, but Brian allows for such coinciding groups in his paper. Even in different possible worlds the groups always have the same members, but they diverge in other features. Microsoft is part of S&P 500 index, but the Microsoft-feature-group is not.

Brian also doesn’t worry too much about parsimony, so he should not have a problem with the increasing number of groups. Having two coinciding groups does not undermine the distinction between types of groups.

Brian could try to argue that the Microsoft-feature-group has the same structure as Microsoft. Organised groups are individuated by their structure, so if the Microsoft-feature group had the same structure as Microsoft, it would be identical with Microsoft.  At least in her 2015 paper Ritchie does not provide identity conditions for feature groups, therefore the argument cannot run the other way. Brian must show that the Microsoft-feature-group has the same structure as Microsoft.

The best reason I see for assuming that the two groups have the same structure, is that they coincide at all times. However, I don’t think that he wants to commit to the claim that if two groups coincide they have the same structure. If the Supreme Court coincides with a golfing club, does the Supreme Court have the structure of the golfing club and vice versa? Probably not. (Admittedly Microsoft and the Microsoft-feature-group coincide over all possible worlds, but I don’t see why that makes a difference.) So, the Microsoft-feature-group could lack all functional structure, although it coincides with Microsoft. At least Brian would have to give us a different reason to think that the Microsoft-feature-group has the same structure as Microsoft.

We have two groups of two distinct types. If I’m right, this challenge to Ritchie’s account fails.