2008-06-23

Do Assume

As the popular phrase has it:

"Assume makes an ass of u and me"

Or does it?

Well I suppose it does, provided that either (a) you don't know the meaning of the word "assume" or (b) you haven't been taught how to assume.

An assumption is not a guess. It is a reasoned or justified statement of some fact that you think is true, AND WHICH you expect to be confirmed or denied before you proceed. The saying ought really to be:

"Acting on what you assume prior to receiving confirmation, makes an ass of u and me."

There are three distinct levels of certainty that seem very seldom to be properly used, and even less often properly understood when they are used correctly:

assume indicates a proposal or question that needs to be confirmed , OR is an explanation for other choices made.

presume indicates that I have some concrete reason for believing something to be fact

expect indicates that I am sure that something will happen

Example

"Arriving at a signpost with a picture of a bus on it, I assumed it was a bus stop and began to wait."

I may be right or wrong.

"I presumed the bus that I saw arriving would stop at the bus stop."

If my first assumption is right, then it is reasonable of me to think that the bus that I saw would stop. Of course, if my assumption was wrong...

"I expected to be allowed on the bus."

By this time the bus has stopped, it's going where I want to go, there's room aboard, and I have money to buy a ticket. I have every reason to think that I will be allowed on.

While expectation is not certainty, it is an honest substitute for it in situations that are too complex for certainty.

"I'm certain that if I drop this rock, it will fall to the ground."

2008-06-17

Distant Precedent

Arguments over style, grammar or punctuation may well make appeals to precedent. This is a good thing if the precedent is recent, since in such cases, it shows current, or at least recent, use.

When the precedent takes the form: this was first used in 1640 (or almost any other date 100 or more years ago), I describe it as distant precedent.

Distant precedent often sounds very authoritative, and many people accept it as settling an argument. This is bizarre. Grammar and (especially) syntax and punctuation can vary enormously over time; it is entirely possible for a form to go in and out of acceptable use over a period of centuries. What matters, therefore, is what is acceptable use ''today''. (see [[less vs fewer]]).

Distant precendent becomes valid only when ''continuous use'' can be shown from the distant precedent to the present. This is a strong justification, but still doesn't trump the [[ultimate rule]].

2008-06-13

Thee Grate Reveler

In Gibbon's indignant account of the destruction of the Pagan temples by the Christians under the reign of Emperor Theodosius, he repeatedly uses the phrase "levelled with the ground" - so frequently indeed that it seems to have already been a cliché when Gibbon was writing.

Today, the phrase has become "levelled to the ground". This is possibly because of the existence of the verb "to level", and the frequent late middle ages description of Death as the great leveller. (I'm told this description was often seen on common depictions such as in Almanachs and even on MA13 of the Tarot de Marseille.) The original cliché seems to have been confounded with "burned to the ground", resulting in the grammatical nonsense of the modern idiom.

The abstract correctness of Gibbon's version makes it rather more acceptable, modern idiom notwithstanding. "levelled to" deforms either the dative pronoun by giving it a genative meaning of (relative to), or deforms the verb, transforming "making level or flat" into "destroy".

I'm going to see if I can slip "levelled with" into the conversation at some point and see if it passes unremarked.

2008-06-09

Naming of Parts

Poem by war poet Henry Reed

An essential first step when learing how to use a tool, and therefore the very first step when describing a tool. Seldom is enough care taken by designers and manufacturers in the naming of parts. If they took a little more care, the job of the technical writer would be so much easier.

The Get Rule

The Get Rule is a PlainWords* rule. The principle is to use the verb to get wherever possible. The purpose is to minimise the amount of vocabulary that the reader needs, and the only criterion is that the context is sufficient that the meaning conveyed by 'get' is sufficient to cover the meaning conveyed by the (possibly more precise) word that get is replacing.

hence:

Getting data

is better than

Retrieving data

since retriving is an obscure latinism, and has no more meaning than 'finding and bringing back' or 'fetching' both of which are in the (rather special) meaning-space** of get. In this specific example, loading data might be better than either, because it is domain-specific***.

Get is a very special animal indeed; it is a metasyntactic variable verb - essentially a placeholder sound whose meaning is conveyed by the surrounding context. What makes it special is that its possible meaning is restricted to notions of transfer. It is the direct equivalent of the latin verb ferre (fero ferre tuli latum) from which we get both transfer and translate.

* "The Complete Plain Words", by Sir Ernest Gowers, is one of my main inspirations
** the linguistic equivalent of phase-space, meaning-space is described as the set of all possible meanings of a given word, excluding identically spelt homonyms
*** referring to the domain of a technical term.