Historical change in comparatives

[Richard]

In the Linguistics course, we looked at whether there had been any change in how comparative adjectives are expressed over time using Google NGrams. Comparing 'more + Adj' with 'Adj-er', we found that there seemed to be a trend for greater use of the bound morpheme -er ('friendlier' became more common than 'more friendly' in 1986; similar patterns for 'quieter', stupider', 'smellier', 'busier', 'cleverer').

However, some adjectives buck this trend: common, pleasant, polite, subtle.

Looking at superlatives, 'pleasantest' and 'more pleasant' cross over each other twice.

The full picture, then, seems quite complex. Can you identify any patterns in this data, or suggest any reasons why comparatives may be changing in this way?

[punjaporn]

I searched COCA for the use of ‘friendlier’, ‘quieter’, ‘stupider’, ‘busier’, ‘cleverer’, ‘simpler’ and ‘likelier’ across five genres (spoken, fiction, magazine, newspaper, and academic). COCA suggests that they are most frequently used in magazine. Is the greater use of the ‘er’ forms influenced by the nature of this media genre which became more casual? But, I still can’t find a reason why common, pleasant, polite, subtle do not follow the same trend. Any ideas?

[Richard]

Aum's idea to check use across registers is an interesting one that I hadn't thought of. Given that the NGram data comes from books (nonfiction I think), does this mean that book authors are writing more like magazine article writers?

To try to explain the differences I was focusing more on phonology, with -ee sounding endings to words more likely to use "er" suffixes, but this doesn't explain all of the words.

[sgtowns]

The full picture, then, seems quite complex. Can you identify any patterns in this data, or suggest any reasons why comparatives may be changing in this way?

I tried a lot of different ideas as to what might be causing these results, but they all failed. I couldn't find any patterns at all. But I did think that complexity or lack of patterns is because these adjectives are all relatively infrequent. None of the adjectives listed (quiet, stupid, smelly, busy, clever, common, pleasant, polite, subtle) are in the Top 50 most common adjectives, at least according to https://www.ef.com/wwen/english-resourc ... djectives/

More common adjectives are all very fixed in their use of more/er and most/est. They are very "black and white". For example,

• More likely to use er/est: new, big, great, high, old, sure
• More likely to use more/most: different, important, right, real, possible, recent

This has not changed much at all over time. So, based on this (very limited) evidence, do the more common words have a "hardened" usage while the less common words are more flexible over time because we have less exposure to them, and therefore seem more "random"? And interestingly, for the common words, the ones that use more/most are either multi-syllable and/or end in 't' or 'l'. Coincidence?

But to the question about what causes them to change over time, I am sure it is a complex issue with lots of factors involved. For example:

• It seems reasonable that phonetics plays a large role, considering the evidence above.
• Aum's idea about genres and formalities might also be a factor. "Pleasant" seems much more formal to me than "smelly". Does that change the way it is used?
• I was also thinking that maybe the relative usage trend of the base word might affect it. For example, "pleasant" peaked in the 1860s and has been on the downslope ever since, while "smelly" appears around 1890 (at least in books) and has been on the upswing to now.
• Or perhaps it could be affected by the changes in the words that the adjective correlate with. For example, according to COHA, the word "pleasant" correlates with "chat" and "weather" in the 1800s but more with "surroundings" and "surprises" in the 1900s. Does that affect other correlations?
• Or perhaps there are differences in whether or not "more" and "most" are used as adverbs or determiners and whether that changes over time too. "Most quiet children will become quiet adults" doesn't mean the quietest children. But Google n-grams doesn't make this distinction.

Seems like we need a lot more data than just what Google N-Grams gives us.

[Richard]

Wow, Stuart's list is pretty impressive. Does it suggest that Google NGrams may be better for initial explorations that identify topics for research, rather than being the main data that is focused on in the research?