Two objections: everyone agrees on what you should do: brush your teeth, be polite, get your five a day, study hard, go to medical school, stop wasting your time on chess etc etc. Doesn't mean everyone does it; doling out good advice is free, but in the real world, one has to cater to individuals' proclivities. Unfortunately, some people want to make everything black and white, whereas life (and perhaps even chess!) is better characterised in shades of grey.

Secondly, and more specifically, I could quote dozens of highly respected sources arguing that ambitious youngsters should study the classics and study endgames. And yet the kidz go their own way, with impressive results. 

So, yeah, Open Games good, categorical thinking bad.

Getting the best out of a student is largely about interpersonal dynamics. Otherwise you may as well just tell them to read a book. So 'my way or the highway' isn't likely to get the job done. It worked for the East German athletics program, no doubt, but times change.

For me, from what I remember, I did openings later when it cropped up, not right at the beginning, when students got stronger, when it became necessary. 

Otherwise:
lots of tactics 
lots of Morphy type hacky games - bits out, forward!
lots of Reinfeld type master-beats-amateur = 'clear plan' games 
essential endgames

Agree with Zwzz about language by the way, well, with regard to language teaching fashions anyhow.

Some opinions I have found on the Open Games vs Closed games debate.

"A beginner should avoid the Queen's Gambit and the French Defence and play open games instead! While he may not win as many games at first, he will in the long run be amply compensated by acquiring a thorough knowledge of the game." GM Richard Reti (!), "Masters of the Chessboard" (Bell,1933)

"Players of the 4th and 5th categories are recommended to play as far as possible OPEN GAMES (except for the Ruy Lopez). It is useful to employ gambits. (...) Here it is very useful to analyse for example, games of Anderssen, Morphy and Tchigorin.
GM Alexey Suetin, "Modern Chess Opening Theory", (Pergamon,1965). 

"There are two ways of solving this problem. One is to select some half-open system and study its hidden possibilities thoroughly enough to remove any fear of surprises that your opponent might have prepared against you. Granted a certain rationale in this method, which does enable a player to achieve perfectly adequate competitive results, a more purposeful approach (from the point of view of improving your play most rapidly) is to opt for 1...e5! In this case it often happens that from the very first moves the game enters the realm of tactical complications, the analysis of which will quickly contribute to raising your class of play. Admittedly this entals much more work."
Vladimir Zak, "Improve our Chess Results", (Batsford 1985). (Zak was an important coach in Leningrad for many years; he was first coach to Spassky, Korchnoi and others. For a dissenting voice, not to say a hatchet-job on Zak, see Yermolinsky's "The Road to Chess Improvement".)

"It is always best, when learning anything, first to try to do the simplest things well and then to learn the more difficult, and this is particularly true here. The objects and methods in closed positions are far more subtle and hidden than those in open positions, and very considerable knowledge and experience, as well as ability, are necessary to play the close game even reasonably well. In the open game, however, both the object - direct attack on some point in the enemy game, often the king's position - and the method - rapid and effective development - are very much simpler and can be understood and reasonably attempted by comparatively weak players. And it is only by starting with the open game that a chess player will learn how to make combinations, an obvious essential to anyone who wishes to play well. Position play will come later and is in any case useless without combinative ability, since it is only by a combination that positional advantage can normally be turned to account." 
IM Hugh Alexander, "Chess", Pitman 1968. (In the period 1935-1955, Alexander was England's strongest player, who took many international scalps, including Botvinnik and Bronstein. His chess career is said to have been constrained by the fact that he was an important figure in British Intelligence.)

"All chess openings starting with 1 d4 are called 'closed' or 'semi-closed'. To play them well and efficiently, one must have a certain understanding of position and some experience. Therefore I advise you to play, at least for a year, 'open' games, and only after this can you start playing 'closed' games."
Garry Kasparov, "Kasparov Teaches Chess", Batsford 1986.

So, except for Yermolinsky, the expert views I have found so far are very much in favour of what has become known here as the "Markovich Doctrine" (cheers, Mark  )

As well as the open games - closed games debate there's another related dichotomy: clear goals vs flexibility. 

Flexible openings (e.g. Flank Openings, the Modern defence, the Nimzo-Indian...) make great demands on a player's positional skill and experience (aka pattern recognition), not only requiring a wide knowledge of different structures, but also the ability to distinguish between favourable and unfavourable versions of the same structure.

GM Nigel Davies has just written about this issue at his website http://chessimprover.com: I quote:

"The basic problem is that (...) flexible openings make great demands on someone’s positional play because they can resolve themselves into many different types of pawn structure. (...) So for a relatively inexperienced player or one whose positional play has not been highly developed it’s much better to play something much more direct in which the position of the central pawns is decided early on and the plans become relatively clear. The London System tends to be very good for this as do the Colle-Zukertort and Torre Attack, and after 1.e4 there are good choices in the f4 Sicilian and Scotch. And if you’re under 1400 you don’t really need to think about these things at all, just get your pieces out and try to assimilate an understanding of a few classic game plans!

One of the reasons more sophisticated openings tend to be popular with Grandmasters is BECAUSE they are so difficult for less experienced players to get to grips with, so they can outplay their opponents much more easily. When beginners play such lines they are not doing themselves any favours, the structural and planning issues will be too complex until they’ve mastered simpler concepts. And they’ll never master the basics whilst wallowing in a sea of flexibility."

Well, I guess Markovich and I would take issue with Davies's suggestions of 1 d4 openings for juniors or inexperienced players, but Davies's main point is sound and important in the context of this discussion.