A steel rod or bar, about 3/8 to 3/4 inch in diameter/thickness and a little longer than the width of the fork, inserted with the straight tines on one side and the "bent" tine on the other - in the direction you want it to be bent, can be wedged down between the tines until a proper common degree of bend is achieved. You will need a way of holding the fork, probably a fairly large hammer, and likely a suitable "drift" punch to drive the bar far enough to be effective. Because of the springback, a degree of "overbend" be required, which can be determined by trial and error. Since you'll probably have to "tap it in" you can proceed slowly and guess at what the result will be. With the bar wedged in place, with a large enough hammer you may be able to additionally "persuade" the tine in a wanted direction by tapping near the wedge to "help it along."

With the wedge to support the mangled tine, tapping between the wedge and the base of the tines will allow you to "localize a bend" that will probably be closer to where you need it, and you'll have less springback to toss the hammer back in your face should you decide to "really whack at it."

There is a possibility that you'll bend the other tines a bit with this method, but there's no "sacred degree of curvature" really needed to make the fork functional. You only require that the tines align reasonably with each other.

The method has been used around the farm for other "tinish" devices, and works quite well with hay forks (where a wood 2x4 may be an adequate wedge). It should be suitable here, if the deformity is reasonably uniform.

Then you can consider the possible ways of getting the bar out from between the tines - which may require more inventiveness than the actual straightening. If the rod used as a wedge is sufficiently long to stick out on both sides of the fork, tapping it out with the same hammer you used to wedge it in quite probalby will not be too difficult.

Fortunately, this is a method with which one can easily experiment, using an ordinary table fork and a butter knife as a wedge. Any one who can bend a spade fork should have no trouble mangling a throw-away dining fork, and the experiment will give you some indication of whether your manual coordination may be adequate to handling the larger garden fork.

I would not recommend that inexperienced persons try heating for a tool of this kind. Unfortunately, the heat and quench method suggested above will most likely result in a brittle end condition, for most metals likely to have been used for a garden fork. The next time you apply a significant load, the tines are likely to break rather than bending. (Breaking a tool is much more hazardous than bending one.)

If you must, a "sometimes" method for tools would be to heat to straw temperature, quench in water or oil, and then place the part in a pan of waste engine oil and light the oil. Let the oil burn itself off, and let the part cool without disturbing it. The flame temperature of the burning oil, in the absence of forced air, should approximate a temperature that will "temper" most carbon steels to a sufficiently tough state to avoid brittle failure the next time it's loaded. Success will depend on the specific metal present, but this method of quench and temper is used occasionally by some gunsmiths I've known for making special-purpose springs when exact replacements aren't available, and usually works well enough with steels somewhat similar to what I'd expect in a common spade fork. It works best with thinner material sections, but should be adaptable to something like your fork. (Expect a lot of smoke while the oil burns off.)

Quite obviously, replacing the fork is probably cheaper, safer, and will get you back to useful work much more quickly. You can cut the tines off and use them for tent pegs at the next campout, if you really feel compelled to do something creative with the old one, but you can expect to consume about one new hacksaw blade per tine doing even that, if the fork is good material.

John