[sent to conlaw prof re: John Austin's Sytem of law in Lecture 1]

Here's the basic problem. Americans tend to talk differently about "law" because their traditions (American constitutionalism) are different. So when legally-educated American eyes see someone saying in print "this is law, but this isn't," the mind sees this stark barrier ("this is what counts, that doesn't").  But an English mind traditionally speaks more colorfully about "law" -- God's law, nature's law, man's law, positive law, scientific law, customary law, the law of fashion, the law of nations, "the constitution in the heart," and so forth.  One might say of the English that they needed to impose a grammar that separated "legal laws" from "non-legal laws." And hence comes Austin who does just that using the tradition of analytic philosophy (although admittedly a little early for what we normally call that). His goal is to impose order. He's trying to say, (a) what is properly called "law" (as opposed to just colorful talking); and
(b) which of this properly-called stuff COUNTS.

To accomplish this, he sets up a pedantic taxonomic system. The system RANKS law. There are four categories to the system. They are

Type 1 -- properly called law, commands from a sovereign
Type 2 -- properly called law, commands from a non-sovereign
(type 3) -- improperly called law, considered "close analogy" to Type 2 (international law, morality, and so forth)
(type 4) --  improperly called, law, considered remote analogy to Type 2 (the laws of physics)
God's law -- properly called law, but not from humans.  

He then takes Types 2 and 3 and dumps them under the heading "Positive Morality" and leaves Type 1 all by itself under the heading of "Positive Law." This allows him, in essence, to imposing what would be considered today as a contemporary American mindset. We now know what's in the bag and what's not:  "God's law" is over there out in left field, the law of physics is out there in right field, the positive morality is in center field -- and that leaves good old  "law's law" in the store all by itself.   

Take a look at a graphic that shows this: http://ludwig.squarespace.com/austin

Now the question is very simple. What did he mean to go into Type 2 that didn't go in "close analogy" [what I have called (type 3)]? The answer to this seems clear:

1. It CAN'T be customs (brithishly constitutional or otherwise), because those are already in "close analogy." The Standford Encyclopedia supports me here. See http://plato.stanford.edu/entries/austin-john/

[Key portion: "Positive law should also be contrasted with “laws by a close analogy” (which includes positive morality, laws of honor, international law, customary law, and constitutional law) and “laws by remote analogy” (e.g., the laws of physics). (Austin 1995: Lecture I)."]

2. He SAYS just a few pages later what a law or rule is. The example he uses is a command between a father and son. This "counts" as a law (therefore must be Type 2). He writes:

"And secondly, a command which obliges exclusively persons individually determined, may amount, notwithstanding [Blackstone’s criteria], to a law or rule. For example, A father may set a rule to his child or children: a guardian, to his ward: a master to his slave or servant.  … Most, indeed, of the laws which are established by political superiors, or most of the laws which are simply and strictly co called, oblige generally the members of the political community, or obliges generally persons of a class."
 
Therefore, father commanding child = law = type 2 = bingo. (Interestingly, I  wonder myself whether your Catechism example wouldn't also be Type 2 to a devout believer who is being directly regimented).