The Article said:
Experts say it's impossible to calculate the number of words in the English language, which is complicated by the classification of compound words, verb forms and obsolete terms.
But don't tell that to Payack.
“We believe words can be counted if you define them in the right way,” he told the Chronicle. “You can count them like anything else in science. You can count how many atoms there are in the ocean.”
This guy is more statistician than he is linguist. (Well, there is a branch of linguistics actually studying more mathematical phenomena than the languages, I believe it is called "Structured and applied linguistics".)
Well, you wouldn't/cannot actually
count the atoms in the ocean - but you can
estimate the number, and you might actually come pretty close - but still, you cannot
count them to check if you estimated correctly.
Another story with the words. You have to actually
count them, you have to have them all in your database to tell how many of them you have - and for that, you do have to define them - whether you count all of the "compound words, verb forms and obsolete terms" (and if they are obsolete, are you sure you know and counted all of them?); you have to decide whether you count dialect words, jargon, slang, etc. That's the part where you have to define what "a word" is, make all the decisions, collect them all (and looks like that's what experts have a problem with - how to make sure that those words are ALL words, no more, no less) and, yes, count them.
Now the second part comes into play. When you say that one millionth English word will be created on June 10, at 10:22 a.m., you are not counting, you extrapolating, or, in other words, you predicting based on estimation made using some previous data. And that previous data also has to be scrutinized, as for criteria on selecting and nominating the new words to be counted.
So in a way, IF you can strictly define words to be counted, IF you can make sure that you actually collected ALL the words that qualify and that you were consistent doing that, THEN you may be able to COUNT them. IF also you can strictly define rules for nomination of new words AND you can make sure that you collected ALL of the qualifying new words, THEN you can observe if there are any tendencies in their creation. Under all of those assumptions, the you may be able to ESTIMATE and PREDICT when a word #1,000,000 will be created. You cannot actually COUNT it BEFORE it is CREATED.
So - not complete nonsense, but not complete truth, either.
Let's say, it is based on many assumptions, which you may or may not believe can be true.
It looks like experts do not believe in the assumptions - so of course they cannot believe in the prediction.
I, personally, is not sure. But then, I do not believe I have the qualifications to rule this case.