The issue is which of them will have observable effects. A rough potted history is that people used to say things like "trait X was selected because it benefits the group even though it costs the individual", but then it was shown mathematically that although that effect does occur, it's very