Pushmi-pullyu exercises

A favourite language-teaching exercise is to make sentences using a construction you're studying. Unfortunately, I really struggle with sentence-making exercises. It's not because of the grammar particularly, it's just something I find very difficult in general, like when people ask you to "say something" - my mind goes blank.

The reason I'm bringing this up is, as you might have guessed, that they make frequent appearances in the textbook I'm currently using.

Basically I think the trouble with these exercises is they test two things at once:

• Your understanding of the grammar point
• Your ability to conceive a sentence that:
  1. Uses the grammar point in question
  2. You know the vocabulary for
  3. You know any other necessary grammar for
  4. Makes sense
  5. Isn't completely banal and self-evident (not strictly necessary, but generally it seems people's minds revolt at saying things that are tautologies, or really really boring)

Part of the problem is that they're not realistic reflections of language. In real conversation, or writing, or reading, you do not spend any of your time thinking of ways to use specific sentence structures. You think of something you want to say, and say it. Language training should guide you towards being able to do that in a different language.

Good language training will encourage you to produce increasingly complex bits of language, assembling blocks together into useful communications. You begin with scattered bits of vocab and grammar that can pass on only simple things, but built up. You learn to convey positive and negative forms; to discuss the past and future; to express hypothetical events; to disparage or praise; to add nuances like amusement, surprise, regret and disapproval; to be polite, or indeed impolite. You go from "The cat is black" to "To be honest, I think I'd prefer that the Tories hadn't implemented a benefits regime that disproportionately penalises the worst-off members of society, but who am I to question the wisdom of a bunch of Old Etonians who've never seen a day's hardship in their lives?".

Pushmi-pullyus do not reliably exercise your sentence-building muscles. This is because completing a pushmi-pullyu exercise requires you to adhere to a number of criteria that, crucially, are often contradictory to a learner. You must create a sentence out of whole cloth, which is difficult at the best of times; but the scope of that creativity is crippled by the need to include a specific grammar point. Many of the examples you might think of require vocabulary you don't have, and which may be actually pretty unusual, and not really worth looking up: "I beseeched the octagenarian cosmonaut to fricasee only the rind of the cantaloupe", for example.

In other cases, a sentence may be perfectly grammatical but not really mean anything, or seem silly, which is a different kind of error and unrelated to the point of the exercise - nevertheless, most teachers will spring on this like leopards upon a hungry gazelle, unwilling to accept the argument that it doesn't affect the validity of the grammar. "If you eat this cake, the giraffe will implode" is a perfectly good sentence; requiring the student to give a detailed explanation of the exact circumstances in which it makes sense to utter those words will not provide any substantial benefit, assuming that they do actually understand what they wrote.

As a result, these exercises can become a frustrating time-sink. It's like the classic scene in films where a fireplace or bin gradually fills with rejected drafts. The student is pulled between different objectives, struggling to make any decision like the ass between two mangers. Ideas are conceived, and rejected. By the time a sentence finally passes the series of disconnected tickboxes, it's taken far longer than most other kinds of question of equivalent length, without any substantial benefit that I can see.

There are at least two obvious ways around this:

1. Offer a set of components to build the sentence around, rather than expecting the student to invent them. Say you're practicing conditionals: offer up "cat", "flap", "open", "come in" and a student can reasonably build "If the flap is open, the cat will come in" or one of several variations, practicing expressing that kind of thought without wasting brainpower trying to come up with the situation themselves.
2. Separate out exercises. Have completion exercises to practice grammatical forms, and free-writing exercises to use creativity (these still need structure - "write something" is deeply unhelpful).

I think many authors and teachers see "write your own sentence" as fun for the learner, because it offers a chance at creativity. But that appearance is deceptive; there are too many constraints for these to be creative. They also seem challenging, but most of the time what they are is difficult.

Okay, that sounds like nonsense. Here's I'm distinguishing two things that both mean "hard" in a way that I'm mostly drawing from computer gaming. Challenging, as I'm using it here, means that something calls for skill to do well, and helps to foster that skill by exercising it. Difficult simply means that something is hard to do.

In a game, a challenging section may require that you execute a large number of moves you have practiced earlier in the game, with good timing, in response to the actions of the robot postmen trying to stop you reaching the dentist. A difficult section often means you must execute a series of jumps from pixel-wide platform to pixel-wide platform, with any mistake resulting in your immediate death, and that some of the platforms are secretly illusions. Difficult parts of games often require luck, memorisation of what is and is not safe to do, or tactics that seem inappropriate to complete them, regardless of the player's competence at the game.

Similarly, I think pushmi-pullyu exercises like "write your own sentence" tend to be difficult because they require simultaneous completion of a number of unrelated and arbitrary goals. It does not matter very much how skilled the student is at forming the past perfective, or pluralising nouns, or using the vocative; the sentence will be difficult to write regardless. In many cases, a native speaker will also have a certain amount of trouble producing a sentence that works within the given bounds.

I'm not saying nobody should ever be asked to write a sentence in the target language, or even that nobody should be asked to write sentences using particular learning points; I'm just suggesting that they deserve a long, hard look and consideration of just how useful they are in building the skills intended at any given point.