> Why don't languages give tail calls their own syntax rather than making it an optimization?

Clojure does exactly that:

    (defn gcd [x y]
      (if (zero? y)
        x
        (recur y (quot x y)))

You can also recur with TCO to an explicitly labelled point within the function (effectively creating an inner, anonymous function):

    (defn factorial [n]
      (loop [n n, fac 1]
        (if (zero? n)
          fac
          (recur (dec n) (* fac n)))))

As I understand it, this syntax was developed because of limitations of the JVM w/r/t TCO.