Historical change in morphemes

[admin]

FROM: Richard Watson Todd (01/28/14 4:34 PM PST)
SUBJECT: Changes in free and bound morphemes

In the Linguistics class, we played with putting variations of morpheme use in Google Ngrams to see whether there had been any historical changes. Examples included:
more stupid v. stupider
more friendly v. friendlier
As a general rule, it seems that the variant with the bound morpheme (i.e. stupider, friendlier) is growing in use at the expense of the variant with the free morpheme.

Three questions on this:

1. How many words/phrases would you need to test for the hypothesis that morpheme use in English comparatives and superlatives is changing to be confirmed? Would you also need to use sources other than Google Ngrams (e.g. COHA)?

2. If the hypothesis is confirmed, can you posit an explanation for why this is happening?

3. Does this pattern of change hold for other linguistic features where there is a choice between use of free and bound morphemes?

[admin]

FROM: Richard Watson Todd (01/28/14 7:52 PM PST)
SUBJECT: RE: Changes in free and bound morphemes


Having quickly played a little more with Google Ngrams, if we want to compare the frequency of use of two competing forms, it's probably better to use the Ngrams syntax and, for instance, search for:

(unlikely / (unlikely + not likely))

Try it!

[admin]

FROM: Pattamawan Jimarkon (01/28/14 8:58 PM PST)
SUBJECT: RE: Changes in free and bound morphemes

I used Google Ngrams with the three following pairs.

cleverer, more clever
prettier, more pretty
classier, more classy

It's inconclusive!

[admin]

FROM: Stuart Towns (01/31/14 11:14 PM PST)
SUBJECT: RE: Changes in free and bound morphemes

Ajarn Pattamawan's data gave me an idea, so I tested it out with a bunch of adjectives that ended in -er and -y and found three different and interesting patterns. I can't embed pictures on Nicenet, so click on the links to see some example patterns in a new window/tab.
1) steady large rise from 1800-2000: friendly, angry, stupid

2) large rise to around 1900 (more or less), then relatively flat: ugly, lively, chilly, silly, lazy, jolly, witty, empty, dusty

3) a small rise to around 1900 (more or less) and a slight fall since then: bitter, sober, clever, tender

Is it just coincidence that the word endings in these three groups are similar (especially #1 and #2 compared to #3)?

So, based on this data, our hypothesis that "the variant with the bound morpheme (i.e. stupider, friendlier) is growing in use at the expense of the variant with the free morpheme" should perhaps include a disclaimer: "except for the comparative and superlative of base words ending in -er."

[admin]

FROM: Richard Watson Todd (02/04/14 5:04 PM PST)
SUBJECT: RE: Changes in free and bound morphemes

It's not just adjectives ending in -y that follow a pattern of a rise to 1920. The same also holds for freer and freest. Is this a phonological issue then?

[admin]

FROM: Steve Louw (02/07/14 3:21 AM PST) [ Send a personal message to Steve Louw]
SUBJECT: RE: Changes in free and bound morphemes

I tried this with COHA. I focused on angrier/more angry because it seems to have the most obvious change of the items that Stuart Ngrammed. I could not actually compare the two items on COHA because comparisons have to be with equal-sized terms.

COHA supports what Stuart found on Ngram. The use of more angry has sort of stayed more or less stable, but the use of angrier has really risen dramatically from around the beginning of the century. In the decade starting 1890, for example, only 9 instances of 'angrier' were found in COHA. The decade of 1900 saw 15, and by 1970 it was up to 30.

I tried to see if there was any pattern in the source of these differences. There doesn't seem to be one. In almost all cases, both angrier and more angry are found in fiction and magazine sources. The only difference I could spot is that with the recent rise in the bound morpheme form, there are more instances in newspapers, which didn't occur in the occurrences prior to 1970. Perhaps the expression of emotion has become increasingly newsworthy, and the newspapers have an inclination towards the simpler form.

My hypothesis, then, is that the 'more' morpheme is not decreasing in being dropped, but that the bound morpheme is increasing in frequency.

I checked this with freer/more free. There has been a decrease in the use of 'more free', and then a fairly steady but relatively low occurrence of it since the 1910s. Freer has steadily increased in usage since the 1850s, and then seems to have fallen from favour again in the 2000s. Again I can't see any pattern in its usage. Unlike 'angry', 'free' seems to be used all genres in both forms.

So my COHA investigation confirms the increase of the use in the bound form. But there isn't any immediate indication of a difference in the reason for this change from the sources.

I then compared both lists from 1890. There are 80 instances of 'freer' in the 1890s, and 35 from 'more free'. Interestingly, 'more free' occurs in written dialogues quite frequently, and in text that is less formal. 'More free' is more associated with poetry and what seems to be legal letters. This might indicate that the bound form is less formal. I did the same with the 1970s, but the numbers are so skewed (freer=60, more free=20) it may not be fair. Both forms occur in "Time" magazine. And of course, 'free time' was one of the collocates with 'more'. It seems we don't yet say 'freer time'

I haven't investigated the phonological hypothesis as a possible reason for the difference.

[admin]

FROM: Richard Watson Todd (02/09/14 10:28 PM PST)
SUBJECT: RE: Changes in free and bound morphemes

Stimulated by Steve's comment about there being no freer time, I wondered how many of the 'more' before adjectives actually apply to the noun after the adjective. To start off with, I use NGrams to compare the uses of comparative adjectives followed by 'than' to all comparative uses:
Comparatives with than

Generally the lines in the NGram graphs look similar.

So then, I decided to do COHA searches using the KWIC function for 'freer' and 'more free'. Generally, 'more free' tends to be followed by 'from' and 'to' with 'freer' being followed by nouns (except 'time'). Interestingly, 'trade' tends to be preceded by 'freer' not 'more free' - thinking about this suggests that there is a possible meaning distinction here.

[admin]

FROM: Stuart Towns (02/14/14 12:44 AM PST)
SUBJECT: RE: Changes in free and bound morphemes

Thanks for the interesting analyses with COHA. I think that going back to the first question originally posed, one can never *only* use Google Ngrams. I think it would be difficult to publish a paper that only had Google Ngrams as a data source. So it's great to see some additional data from other sources.
By coincidence, an blog posting from a Corpus Linguist talking about the pros and cons of Google Ngrams popped up in my RSS feeds last week. If this thread has been at all interesting to you, you might also enjoy this article:

http://andreadallover.com/2014/02/09/dont-go-down-the-google-books-garden-path/

P.S. For those taking Linguistics this term, the title of this blog post is a nice usage of the term "garden path" that we talked about a few weeks ago.

[admin]

FROM: Stuart Towns (02/23/14 9:14 PM PST)
SUBJECT: RE: Changes in free and bound morphemes

Here's another interesting write-up about COHA vs Google N-Grams -- but this time from the folks at COHA. It looks like you can also do some limited searches on the Google Ngram "corpus" from the COHA website. Interestingly, they say that COHA and Google Ngrams give similar results on frequencies.
Click the link below for more info:

The Corpus of Historical American English, Google Books, and our new Google Books interface