## Saturday, December 18, 2010

.

### How many are that?

It might surprise you — it surprises me — that number agreement is a serious problem in technical papers and specifications. Having just reviewed a specification that’s full of number-agreement errors, I feel an urge to talk about the problem here.

In the easy case, things are... well... easy:

Bill and his brother are tall.

Bill and his brother is tall.

The second sentence is clearly both

1. wrong and
2. unambiguous.

That is, we know it’s wrong, but we also know what the writer meant to say, and, apart from tripping for a moment before reading further, we have no problem understanding this.

But try this pair:

I saw Bill and his brother, who are tall.

I saw Bill and his brother, who is tall.

In this case, both sentences are correct, but only the first one tells you that Bill is tall; his brother is tall in both of them. Now, it matters that we get the number correct.

Of course, you say. Any fool knows that.

Well, yes, mostly. But when sentences get complicated, some of this gets lost. Let’s try this one:

Data objects have a set of well-defined fields that include a single value and optional metadata.

Look at what we have in the sentence:

1. Data objects. [plural]
2. A set of fields. [singular]
3. Fields. [plural]
4. A single value. [singular]
5. Metadata. [collective; singular or plural]
6. A single value and optional metadata. [plural]

As it’s written above, the sentence says that multiple data objects share one set of fields. That’s not right, so let’s fix that first:

Each data object has a set of well-defined fields that include a single value and optional metadata.

Now comes the question that I don’t know the answer to: Which of the following is correct?

1. The set of fields contains one field with a single value and one or more optional fields that each contains metadata.
2. Each field in the set includes a single value and optional metadata.

The way the sentence is worded, number 2 is correct (imagine parentheses that start before well-defined and end after metadata). But the fact that the document is full of number-agreement problems makes it unclear what the authors really mean. The sentence has to be reworded:

Each data object has a set of well-defined fields that includes a field with a single value and optional fields with metadata. [Here, a set of well-defined fields is singular, so the verb is includes.]

...or...

Each data object has a set of well-defined fields; each field includes a single value and optional metadata. [Here, each field is singular, so, again, the verb is includes.]

The trap we too often fall into is forgetting that a set of things or a group of things, while it might look plural (especially when the clause defining the things is complicated), is singular. For example, compare these:

That group of Tea-Party voters who voted for the Republican candidate in the last election and who are looking for big tax cuts is influential.

Those Tea-Party voters who voted for the Republican candidate in the last election and who are looking for big tax cuts are influential.

Be careful of the difference. It can really matter sometimes.