For example, the amount of moisture in your skin: too little and it cracks and bleeds, too much and you get fungal infections. But adding moisture to cracked-and-bleeding skin isn't making it "a bit more fungal" -- it's pushing it into the healthy buffer range.
I'd imagine there are many, more complex, factors involved in brain chemistry / mood suppression, so it would be a lot more difficult to determine where this middle lies.
If you're using high suicide rates to suggest we're not in the middle and need more lithium in everyone's water, I'd have to disagree. In the vast majority of cases, a better healthcare system which ensures those who need it are connected with mental health professionals would help nip the problem in the bud. Rather than solving it with these lithium microdoses, causing to those on the verge of suicide to not have the lucidity of emotion to commit to the act, and the rest of the population having their brain chemistry played with too.
Here's a useful comparison:
Before we were iodizing salt, most people were actually in "the middle" of iodine-consumption. Anywhere from "nearly none" to "tons and tons" was "the middle"--your body can work miracles to create homeostasis from random inputs. But some people were getting no iodine at all, and these people would get very sick.
So we added just a little bit of iodine to salt. The iodine-deficient people--the people who were previously getting utterly no iodine--stopped getting sick, because now they were getting a little bit of iodine, and that was enough. Everyone else was getting slightly more iodine--but this didn't push anyone from "getting a tolerable amount of iodine with the excess just excreted" to "having a hyperiodine problem." There was nobody exactly one dose away from having too much iodine, because we just don't consume that much iodine.
This same story holds for any vitamin or mineral. Calcium, phosphorous, iron, etc. are all things most people get the right amount of--but "the right amount" is almost any amount, because the body knows how to take just the right amount from the stream of input it's getting, and throw the rest out. The only wrong amounts are either zero, or more than exists in any human-friendly environment.
Actually, there was: 'thyroid shock'. Iodization in the USA was accompanied by a wave of, I think, a few thousand deaths on net when people who grew up iodine-deficient were suddenly exposed to iodized salt. (This was covered by one of the historical papers I cite in http://www.gwern.net/Iodine although I forget which.)
Of course, the benefits of iodization were on net huge and I'm very much a fan of iodization, but it would be wrong to say nobody was hurt by the introduction of iodized salt.
I guess another way to think about environmental chemicals is by analogy to poisons: if they were common in the environment, you'd have a tolerance to them, but since they aren't, you don't. You can build up a tolerance by parts, at which point the poison stops being one and just becomes another chemical the body handles.
What is there to show that "this same story holds for any vitamin or mineral"?! Surely there are many which have no positive effect on the body, where zero would be the best dosage. What law of biology suggests that every mineral is in some way beneficial to the human body?
Lithium happens to have an effect on moods in the brain in such a way that suicide is less likely - but the mood dampening effects may well diminish people's positive emotions too (for example). For someone with no mental health issues or risk of suicide, this could well be an overall negative effect, and therefore makes no sense to force on the entire population.