No true scotsman

I have a problem with the No true Scotsman fallacy. I fail to distinguish it from a modens tollens. Take this famous argument for example:

Person A: "No Scotsman puts sugar on his porridge."
Person B: "But my uncle Angus is a Scotsman and he puts sugar on his porridge."
Person A: "But no true Scotsman puts sugar on his porridge." 

It can be arranged in a syllogistic form and still retains its original form and meaning

If you are a Scotsman (A), then you wouldn't put sugar on your porridge (B)

Uncle Angus puts sugar on his porridge(-B), therefore he is not a Scotsman (-A)

Can anyone help explain this? Thanks

Again, as I incessantly preach, start with the premises. “No Scotsman puts sugar on his porridge.” If this premise is true, then the syllogism works. BUT IF we have to define a Scotsman, then the premise will be clear, and the syllogism falls of its own weight. The error comes partly from reading more into the premise than the premise states. Perhaps this viewpoint makes the fallacy more clear.

The modens tollens (MT) is an argument form and the no true scotsman is a fallacy. Notice with the MT, it is a conditional (has an "if"). The no true scotsman is contradicts a point of fact rather than states a conditional. Uncle Angus is, in fact, a Scotsman who puts sugar on Porridge. Therefore, the error in reasoning is the claim that "no true Scotsman puts sugar on his porridge." It is the attempt to redefine a term on a whim. With the MT, we accept the premise that "If you are a Scotsman (A), then you wouldn't put sugar on your porridge" is true. The form is valid but the argument itself is unsound because of the false premise.

In short, with the MT, the error is in the first premise (with the conditional that is false). With the no true scotsman the error is premise/claim that "But no true Scotsman puts sugar on his porridge."

Your observation that Modus Tollens seems to be related to the No True Scotsman is not misguided. However, the fallacious part of the No True Scotsman is informal and rather than being caused by an error in the syllogistic structure is instead caused by employing Modus Tollens to an ad-hoc redefinition. It's actually quite a tricky fallacy to wrap your head around.

Take scenario (A): If one enters into a dialogue defining from the onset of the related conversation that part of the definition of a Scotsman is "one who does not put sugar on porridge" then one would be absolutely entitled to employ Modus Tollens to prove that Angus is not a Scotsman using the exact syllogism you presented. There would be no fallacy committed.

However, in scenario (B): If one enters into a dialogue defining from the onset of the related conversation that a Scotsman is merely "anyone with Scottish descent", clearly Modus Tollens using sugar will not work unless we do one of two things:

(1) we could prove that the truth-value of the proposition " 'anyone of Scottish descent' does not put sugar on their porridge" is true
(2) we could do some redefining to change our definition from "anyone with Scottish descent" to something like "anyone with Scottish descent who does not put sugar on their porridge" to make (1) trivially true

So, let's apply this to a real example. In our hypothetical scenario with Angus, Person A opened up by stating: "No Scotsman puts sugar on his porridge". Did Person A put forth a proposition with an empirical truth value they are asserting is true or are they simply stating their definition? This is crucial to decide because it is what determines if we are in scenario (A) or scenario (B). If definition, scenario (A) and not the No True Scotsman. If an empirical claim, scenario (B) and we have more digging to do.

To most people, the assumption will be that Person A is making an empirical claim. Person B seems to be under that assumption, since they reply with a counterexample as empirical evidence to the contrary. If Person A had been simply stating a definition, Person B would have no business providing a counterexample because a counterexample to a definition is not possible; definitions are arbitrary (setting aside meta-discussions of which definition is more pragmatic for efficient communication).

So, we now have an empirical claim and an alleged counterexample. Person A's options of (1) and (2) for pre-requisites of employing Modus Tollens are now relevant. Person A could try (1) by sidestepping Person B's counterexample and proving by some other means that Person A's assertion is factual but this would not address why in particular Person B's counterexample doesn't disprove the claim. Still, it's a plausible movement and more importantly it would not be a fallacious No True Scotsman move. Alternatively, Person A could more weakly try (1) by using some outside evidence to disqualify the counterexample. However, it's not even a guaranteed win for Person A because Person B's counterexample could actually be a true counterexample, disproving the claim and making an alternative proof of the claim impossible. Whether or not the counterexample is solid, Person A could try (2) and this would directly exclude Person B's counterexample by Modus Tollens but this would be a committal of the No True Scotsman fallacy. Remember that (1) and (2) are not mutually exclusive. In fact, (1) is a direct consequence of (2) which is why (2) is so tempting.

In Scenario (B) where Person A tries (2), they have first postulated the truth of a proposition with an otherwise "unknown" truth-value but then tried their very best to switch back to Scenario (A) at the first sign of trouble to guarantee that desired truth-value by definition.

They have changed from making a claim that a proposition is true: " 'anyone of Scottish descent' does not put sugar on their porridge" to  stating a definition: "anyone of Scottish descent does not put sugar on their porridge". This is clearly fallacious to change from an empirical claim to a definition in the middle of a conversation. Person A would've been fine to employ a definition and Modus Tollens in scenario (A) but here in (B) they instead made a truth claim that was about to be potentially disproven by a counterexample and only then did they try to change their initial definition so they could use Modus Tollens to show a trivial truth. This new, changed, definition is  ad-hoc because it's only being used to target this one very specific counterexample. Person B has evidence it's ad-hoc because they saw Person A use a different definition before this counterexample was presented at the onset of the related conversation.

When Person A makes the transition, they'll hide their modification with purity platitudes such as "true", "real" to essentially accuse Person B of not using a "pure" enough or "airtight" enough definition. It's akin to proclaiming, "when we both re-evaluate our definitions, we should agree a true Scotsman definition would define a Scotsman as never putting sugar on their porridge!" But of course, Person B should not allow Person A to get away with this rhetorical sleight of hand, because they should not be in the business of re-evaluating the definitions after a claim has been made in the first place.

Hopefully this helps