Last semester in the class I was teaching I did an experiment that I thought had clearly interpretable, interesting results. Well aware that my students were quite unaccustomed to linguistic analysis, I didn’t expect any of them to see all that was obvious to me, but the write-ups I received were still disappointing. Many of the students did no linguistic analysis, and those that did rarely noted even the patterns that I thought were really basic. So here, as catharsis, I will present the experiment and what I conclude from it.

The question we start with is this. Officially, saying something like “Me and my brother were chatting” is ‘bad grammar’. However, people say stuff like that all the time. Also, related phrasings that are officially more grammatical, like “I and my brother were chatting” are rarely used. Why is ‘bad grammar’ (like “Me and my brother were chatting”) so common while other related options are rare? The answer to this question reveals that ‘bad grammar’ and ‘proper English’ aren’t quite what people normally conceive them to be.

I presented my students with the following example sentences and asked them to rate their acceptability on a scale of 1 to 5, where 1 is something that doesn’t really sound like English, 3 is something you don’t think you would say but you hear other people say, and 5 is something you would feel fine saying. (The implication here is that sentences rated 2 and 4 lie somewhere between the sentences rated 1 & 3 and 3 & 5, respectively, but there were a couple students who missed that, thinking the only options were 1, 3, and 5. Do we speak the same language?) We then averaged everyone’s ratings. My deapest apologies to those of my readers who care about statistical significance, since I did not collect the data on a case-by-case basis, precluding variance calculations.

There are several patterns worth noting here. The first sentence (a) was included just to verify that people understood the task, so let’s start with the second block.

In sentences (b) through (e), note that there is a very strong preference for ‘I’ rather than ‘me’ as the subject of the sentence, and a similarly strong preference for ‘me’ rather than ‘I’ to come as the object of the preposition ‘with’. It’s because of this that ‘I’ is often called a subject pronoun, and ‘me’ is called an object pronoun. This contrast between ‘I’ and ‘me’ is very widespread, and similar contrasts are common in other languages and used to be much more common in English. But now, this sort of contrast only appears with personal pronouns. Glancing ahead quickly, also note that none of the other sentences recieve scores as high or as low as sentences (a) through (e). The more complex sentences are neither as acceptable as the good sentences here, nor as unacceptable as the bad sentences here.

In sentences (f) through (i), we note that between (f) and (h) there is a strong preference for ‘I’ to come after ‘my brother’, whereas between (g) and (i) there is a noticeable preference for ‘me’ to come before ‘my brother’. The other thing to notice is that, in contrast to the pattern we noted in (b) through (e), ‘me’ seems quite acceptable in the subject. In fact, if we average the score from (f) with (h) and the one from (g) with (i), we get 3.05 for both! The choice between ‘I’ or ‘me’ has no net effect on the acceptability of this sentence. If we average the other way, we see that there is an overall preference (3.65 vs 2.45) for ‘my brother’ to come before the first person pronoun. To explain this preference, you may recall from grade school an admonission to “put others before yourself” in situations like this as a sign of respect. But why is “I and my brother” still less acceptable than “My brother and me” or “Me and my brother”, contrary to the strong preference for ‘I’ that we observed in (b) through (e)? We leave this question unanswered for the moment.

In sentences (j) through (m), we note that (j), (k) and (l) are all approximately of the same acceptability, whereas (m) is much less. Based on the patterns we have noted above, it is quite understandable that (m) is rated so poorly since it violates two generalizations: we have the subject pronoun ‘I’ in the object of a preposition, and we have it preceding ‘my brother’. Note however, that sentences (k) and (l) each violate one generalization, while (j) violates neither, and yet (j) is the worst rated of the three. What is going on here?

To answer these questions, we will take a look at some usage data in part 2.