- a topic that everybody knows will be the focus, and
- an audience that actively values your conversation.
Outside your core discussion time is when you'll see general discussion tropes. If people are (generally) neither leaving nor arriving, in my experience we see transient circles of 6-8 that split (if more than 8 people, or if multiple topics persist) and merge (if fewer than 6 people). The article's limit of 4 doesn't apply if movement between different groups is considered a feature rather than a bug.What's really interesting about your observation is how the rule-of-thumb breaks down when the conversation is limited to more confined topics or: If the individuals see each other with some regularity outside the group setting (no need to engage in extended conversation about who's doing what) or some of the conversations involve topics brought up in the course of the class.
Care needs to be taken, of course, but ancient works often followed certain patterns or linguistic choices that could be used to identify authorship. As long as this is viewed as one tool of many, there's unlikely much harm unless scholars lean too heavily on the opinions of AI analysis (which is the real risk, IMO).