4

My hypothesis is that in informal writing (say something like Stack Exchange questions) "it's" is massively more common than "its".

Is this true? Are there any data to support it? Concrete percentages?

Is there a heuristic that a very simple program¹ can use to detect when "its" would be appropriate?

The purpose of this question is that it is one of the more frequently confused pairs of words. I'm designing tools for flagging potential spelling mistakes, so if one was a lot more frequent than the other and I had indicators that the piece of text was poorly spelled then it might be a valid suggestion.

| improve this question | |
  • 1
    The second half of this question is an exact duplicate of english.stackexchange.com/questions/653/… The first half seems interesting, though. – PLL Mar 21 '11 at 19:57
  • 2
    Speaking of spelling, you might want to examine your use of "then" vs. "than". Just sayin'... ;-) – Marthaª Mar 21 '11 at 19:58
  • 2
    In common speech, it's and its are indistinguishably pronounced the same. – nohat♦ Mar 21 '11 at 20:00
  • @nohat Ah yes, I didn't exactly think that through. – Jakub Hampl Mar 21 '11 at 20:02
  • 4
    @jae: Sadly, in English there is mixed conception of the plurality of 'data'. It is undeniably plural in the original Latin, but colloquially in English it acts like a mass noun (like 'water': "Is there any water?"). But pedants and scientists tend to prefer the plural interpretation. (one 'fact' in favor of the singular (as if that matters) is that the Latin singular 'datum' does not really occur in English) – Mitch Mar 22 '11 at 0:07
4

I would say that It's is more frequent than Its.

It's = it is

Its = owned by it

Two completely different words, although with similar uses

I think it's generally more common to speak about things using It's when describing them than its as the sentence just feels more natural, for example

It's red vs Its colour is red

As for data to support this I'm not sure, however when I studied Natural Language Processing it is possible to use bigram / trigram equations etc to predict a word given the word / words before it. Probably something similar would work here.

| improve this answer | |
3

In common speech, they sound the same, so the question doesn't apply. For common writing like SE questions, I don't know, but for more formal writing covered by the Google Books Ngram corpus, its is around 8 times more common than it's (graph).

Re a heuristic, I believe one could have a set of rules that would cover a moderate percentage of cases with medium accuracy, but not with high enough accuracy to automatically correct, or probably even to flag for human review, incorrect usages. Deep grammatical parsing is needed to reliably evaluate the cases.

| improve this answer | |
  • That's an incredibly useful link, thank you! Interesting is when you chart to 2008 "it's" goes to be only about 3 times less common. – Jakub Hampl Mar 21 '11 at 20:06
  • 1
    That's because usually you don't use contractions in books. If you look for "it is" the picture is quite different. Also, interestingly, "Its" is less common than "It is" (with the capital I, so essentially at the beginning of the sentence). – nico Mar 21 '11 at 20:08
  • It's simple to add "it is" to the graph: ngrams.googlelabs.com/… – Marthaª Mar 21 '11 at 21:51
1

No, there is no algorithmic determination possible. For example you might think that before an adjective, there should be it’s, but think of: “it’s possible to do this” and “I understand its possible consequences”.

| improve this answer | |
1

  1. I have no idea. Given that "its" and "it's" have completely different meanings, I don't know what purpose such a comparison would serve.

  2. its is the possessive case of "it", i.e. "belonging to it." it's is a contraction of "it is". It doesn't get much simpler than that.

The zebra picked up its suitcase and walked off. "It's a miracle," George said.

Edit: I misunderstood the "very simple program" part of the question, sorry. Unfortunately, very simple programs fare very poorly against English; it has borrowed from so many different languages that most simple heuristics will lead you wildly astray after a while.

Edit to match the question edit:-) Sorry, doing it properly needs a grammar check. See Microsoft Word for the utter howlers that result from trying to do grammar checks on the cheap!

| improve this answer | |
  • That's not a simple algorithm. That's obvious when you look at it, but programming that takes a lot of effort. – Jakub Hampl Mar 21 '11 at 19:57
  • @Jakub Hampl: yes, I realised I wasn't answering your actual question while you were commenting! Sorry about that. – user1579 Mar 21 '11 at 20:00
0

Is there a heuristic that a very simple program¹ can use to detect when "its" would be appropriate?

Yes, but the simpler the program, the simpler the heuristic. The simpler the heuristic, the more mistakes will be made.

For example: From a sufficiently large body of English text, you could build a table of word-pair frequencies involving "its" preceded or followed by other words and "it's" preceded or followed by other words. Your simple program could use this to determine whether a word pair would be much more probable with "its" instead of "it's" (or vice versa). Looking up a pair of words and associated score ought to be something a simple program could do.

| improve this answer | |
  • I didn't <del> it because I believed it was off-topic but because I was much more interested in the other part. I originally wrote this just in case I was missing something obvious like "its is used always after the word foobar". – Jakub Hampl Mar 22 '11 at 15:11
  • @Jakub: OK I'll edit the answer to remove the false implication. – RedGrittyBrick Mar 22 '11 at 15:16
0

I've found the default-to-'it's' in my smartphone's autocorrect feature to be maddening. My perception is that I use 'its' far more frequently than 'it's'. But I suspect that the iPhone simply never recognized 'its' as a correct form. Given my iPhone's decent ability to figure out which other words I'm intending to type, I fear that the developers may not realize that there's a difference between the two words. I had the same problem back when I had a Palm Pre.

But a frequency algorithm would be a step in the right direction. In English, 'its' would not typically precede an article or a verb. Seems like this must be a solved problem somewhere.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.