All energy can be equated to a force acting over a distance. While it's true that some kinds of energy (thermal, for instance), can't be litterally be broken down into a specific force over a specific distance, that doesn't mean that the thermal energy doesn't have an 'equivalent' that is force times distance. We can say this amount of thermal energy is this many Joules... which means it corresponds to this many Newtons over this many meters.
Perhaps a better way of describing thermal energy is actually to break it down into tiny little energies. Put a thermometer into a cup of hot water - what the thermometer reads is basically a statistical sampling of the individual water molecules' kinetic energies summed over the exposed surface of the thermometer multiplied by the energy distribution of the water (modulo the thermal transfer properties of the water and thermometer).
Or perhaps in terms Mr. Rea might find more familiar - thermal energy is an integrated measure of individual particles capacity to do work over a distance.
Consider PV=nRT as it relates to a gedankenexperiment with a rectangular prism shaped chamber filled with an ideal gas. One of the chamber's walls is freely movable. (Frictionless etc. etc.) At STP, a given amount of gas "n" will come to equilibrium so that the volume of the chamber is given by nRT/P.
What happens when you heat the gas?
"Microscopically," the gas molecules hitting the movable wall from the inside are balancing the air molecules hitting it from the outside. When we heat the gas, the molecules closest to the heat source will receive energy from said heat source (generally through collision or radiation) and proceed to transfer energy to other gas molecules (etc. etc.) until the entire collection of molecules come to statistical equilibrium. As this is happening, the now more energetic gas particles are colliding with the movable wall, pushing it outward until a new equilibrium position is reached.
Take the "Macroscopic" point of view. Since the amount of gas doesn't change and external atmospheric pressure is constant, the volume of gas will expand until a new pressure equilibrium is reached.
Same thing (Kinetic energy integrated over each molecule =~ Temperature), same result!
But notice. In both ways of looking at the system, the movable wall moves due to Pressure|Energy Transfer. Does this not mean that the KE/Thermal Energy has a direction that points towards the wall?
No. Consider the same experiment, but with two movable walls. In which direction does the energy point? Or three. Or model a balloon as an infinite set of infinitely small walls.
The answer is that in the first case, the action is in the direction we put the movable wall, because that's where we put the movable wall!
Put another way. In Mr. Rea's example of heating one end of a metal bar, the heat moves from one end to another because that's the shape of the bar. The apparent "vector-like" direction of energy transfer in this case is not a property of the thermal energy itself, but with the vector-like setup of the experiment!
"Hot to Cold" is a 'direction' in the same sense that "Radially Outward" is a 'direction' in the expanding balloon case or the heated disk I asked about elsewhere. It is descriptive terminology, not an inherent property of the energy itself.