Most categories of difference are fairly straightforward: it is obvious that one lion is different from three lions, or a green lion from a red one. The one area of difficulty is difference of type: is a lion different from a domestic cat? an heraldic tyger? a chevron? How different is different enough?
SCA heraldry counts three degrees of differentness. Two charges can be, in order of decreasing difference, (1) substantially different, (2) significantly different, or (3) insignificantly different. The definition of each degree of difference goes back to the concept of cadency, discussed in the first installment.
Two charges are considered substantially different if the change is more than a cadency step. For example, compare a lion and a chevron. As a thought experiment one might imagine a period coat "Argent, a chevron sable." Might a younger son show cadency by changing this to "Argent, a lion sable"? No, this is not plausible, as the change is to great and the family relationship would be lost. Therefore this change is more than a cadency step, and hence is considered substantial. The vast majority of possible combinations fall into this category. As an addendum, different sorts of beasts are usually substantially different: lion vs. wolf vs. horse vs. armadillo. I only add this because many years ago this was not the case, and some old ideas die hard.
Two charges are considered significantly different if the change from one to the other is equal to at least one cadency step. Note that all substantial changes are also significant. The reverse, however, is not true: there are changes which are significant, but too small to be substantial. For example, compare a chevron vs. a chevron embattled. Might the base coat of "Argent, a chevron sable" be modified by a younger son to "Argent, a chevron embattled sable"? Yes, this is a plausible change. In general adding or a complex line of division is a significant, but not substantial change. Changing from one complex line to another is usually, but not always significant. Other significant (but not substantial) combinations include oak tree vs. pine tree, lion vs. (heraldic) tyger, and horse vs. unicorn.
Two charges are considered insignificantly different if the change from one to the other is less than a cadency step. This includes changes in the head position of a beast, e.g. a lion passant vs. a lion passant guardant. Some posture changes fall here, e.g. passant vs. statant, rampant vs. salient. Maintained charges are considered insignificant, e.g. a lion rampant vs. a lion rampant maintaining a sword. Other combinations are a bit trickier. Sometimes two charges are regional variants of the same thing: in Continental heraldry a "dragon" has two legs, while in English heraldry it has four. Therefore we give no difference between them. This is obscured by the fact that English heraldry occasionally uses a two-legged version and calls it a "wyvern". Another example is a mermaid vs. a melusine: while they each have distinctive characteristics they are simply the English and Continental versions of the same idea. Other charges are temporal variants of a charge. In early heraldry the cross crosslet had rounded lobes. In later heraldry its corners were squared off, and the rounded version was re-invented as the cross bottony. The SCA therefore considers the distinction insignificant. Often variations within a type are insignificant: all felines are the same ("cats is cats"). Other examples of insignificant changes include rose vs. cinquefoil, tower vs. castle, and wavy vs. nebuly.
How is a hard-working herald to know which changes count for what? There are several approaches. Often the organization of the Ordinary can be a clue. For example, it lists castles and towers together. This needs to be approached with caution. The Ordinary's organization tends to lag behind current practice. The current Morsulus Herald (the herald in charge of the Ordinary) is doing an outstanding job at bringing it in line, but its history still shows through. My 5th edition lists ermined and vairy fields as field treatments, yet the CoA abandoned that interpretation many years ago. Also, not all Morsulus Heralds have been as vigilant as the current one, and we have no way of knowing what his successors will do. The Ordinary can give a clue about questions of difference, but it should never be taken as the final word, or as giving official policy.
A second source is the Pictorial Dictionary. When it tells us that a wyvern is a Continental version of the English dragon, this is a pretty good indication that no difference is given. Often it will explicitly state whether significant difference is given. (Generally if the question even arises there is little chance of substantial difference.) Again, the Pic.Dic. is not an official document of the CoA, and may be out of date in individual cases. Its reliability is very good, but not perfect.
The third source is finding precedents in previous Laurel rulings. While Laurel can overturn existing precedent, this is rarely done unless new evidence demands it. Generally the more recent the precedent the better. Significant rulings of the various Laurels are collected, edited, and published by Free Trumpet Press West. There is some talk of putting them on-line, but they are currently available only in paper form. For practical purposes few precedents from before Da'ud I still apply, so earlier collections are of primarily historical interest. (The Early Years volumes, by the way, make for fascinating reading!) These collected precedents are how one can find heralds rattling off rulings from ten years ago. (And you thought they had them memorized.)
There is also the question of how do we decide when there are no applicable precedents. This is a decidedly advanced topic, and I am trying to keep this tutorial as short and simple as possible. This is, however, the sort of thing that makes research heralds' juices flow.
In terra pax,
Rouland "What, Me Worry?" Carre