What didn't I understand? He insists there has to be a reason why we cry (as understood by tearful weeping) and spends the rest of the article trying to ascribe a social purpose to it. His rejection that crying is therapeutic is a cornerstone to that, and it ignores a whole range of alternate lines of inquiry.
Exercise has a host of indirect (endorphin production), secondary (improved circulation), and tertiary (self-image) effects that improve mood. I think his failure to imagine other reasons for crying are just as limiting as trying to deny those effects of exercise because sweat doesn't contain mood improving hormones or excrete mood worsening ones.
1) He didn't deny that crying is therapeutic in the emotional sense (in that you feel good afterwards); only that the expulsion of tears or the quavering of breath serves no regulatory function physiologically. Just because it feels good after those things happen doesn't mean they caused it.
2) His point is that the urges to cry and feeling good afterwards are evolved instincts to coax you to do something socially useful, i.e. to signal to your kin that you're in distress. The tears and sobbing are the signal; an effect of your emotional state, not the cause of it. The tears are intended to act on your peers, not your physiology. When they've done their job, you're relieved.
In other words, the "symptoms" of crying are probably arbitrary and could have been any number of things that accomplish the goal of signalling, because there is no reasonable physiological connection to the display and the change in emotional state.
3) Exercise is not a good analogy because it has a direct physiological effect on your system. It is exercise which strengthens your muscles and cardiovascular system, not any other activity. Thus it makes sense that we feel internally rewarded for having exercised, because the activity is directly needed in some sense to maintain homeostasis. Compare this to his hypothesis about crying, which is that if its purpose is truly to be a signal, it would have made just as much sense for us to evolve to, say, shiver, or urinate, or faint, or vomit, or convulse, or do any other highly recognizable behavior involuntarily to indicate distress to our peers.
1) Why don't you think it's physiologically therapeutic, too? As I note in another comment, we have to correct drug studies for the placebo effect because we actually get better if we believe we're going to. Breathing changes can modify mood. Maybe this is incorrect, but I'm certainly not going to accept it serves no physiological function without any sort of proof.
2. I believe you are misunderstanding his point, which is, "inherently they're just a piece of social technology, a device for coordinating the tradeoff between dominance and social support."
2b) You undermine your point #1 by pointing out that the "symptoms" are arbitrary. Out of the thousands of possible responses, why tears + weeping? There are more effective signals, and other social animals acknowledge dominance & submission differently. That would suggest crying has a reason beyond social signaling.
3) I don't think exercise is a poor analogy here, but let me be direct: crying could have a direct or secondary neurological effect on mental state that doesn't have an analog in other animals.
