Linguistics 300, F08, Assignment 9

Due date: M 11/10

In this assignment and the next one, you will be collecting the data that will underlie our investigation of shifting word stress in English. Be forewarned that both assignments are likely to be a bit time-consuming. Please complete your part in a timely manner, as delays in collecting the data will affect not only your grade, but the quality of the database that we have to work with. If you foresee any difficulties with meeting the deadline, please let me know earlier rather than later.

The present assignment focuses on identifying suitable words for analysis, based on the information available in the OED. In the next assignment, you'll be collecting and analyzing verses containing the words that you are now identifying, based on the concordances that you've identified earlier in the class.

The dataset that we are interested in for the purposes of the assignment is the set of French borrowings that entered English before 1625. This terminus ante quem (as they say in philology; literally, Latin for 'the time before which (something happens)') is chosen to roughly coincide with 1616, the year that Shakespeare died, whose writings will provide a major source of evidence. In the interests of completeness, we will not set an early bound on the data. However, we would expect there to be few French loanwords before 1066, the date of the Norman Conquest. Given the entire set of French loanwords, we will eventually focus on those with more than one stressable syllable, since monosyllabic words or disyllabic words with an unstressable syllable aren't candidates for stress shift. For the moment, however, we will exclude only words that are clearly derived from monosyllabic etyma. We will retain words like cattle with apparently unstressable final syllables. The reason for doing so is that these words might be derived from etyma with final syllables that were stressed in (older) French. In other words, the unstressability of the final vowel in English might be related to the very stress shift that we are interested in.

In an optimal world, the OED would be a true database, and it would be possible to submit a query with the conditions specified above (something like etymon="FRENCH" AND range="<1620" AND syllables">1"). Unfortunately, however, the OED is not (yet) a true database in that sense. First, the good news: It is possible to limit the results of searches by dates for first attestation, which is what we will use as the best estimate for when a loanword entered the language. In order to do so, you can either search for a range (e.g. 1066-1100, -1620, 1900-) in "First cited date" or sort the results of previous searches by "Entry date." The bad news is that, although the OED contains information concerning pronunciation, the pronunciations are not searchable, so that we can't limit the results of searches by stress pattern or even by number of syllables. And although the OED distinguishes between cognates and etyma in its etymological entries, distinguishing between these two types of etymological relationship is not completely straightforward.

The etymon of a word is its direct historical source. For our purposes, we are interested in English words with a French etymon. It is also possible for two words from related languages to be related in a more indirect way. To take a concrete example, the Germnanic languages (including English) and the Romance languages (including French) are both daughters of Indo-European. The Indo-European word for, say, 'mother' has a Germanic descendant, which eventually develops in English mother, and a Romance descendant, which develops via Latin mater into French mère. In kinship terms, we can think of the relationship between mother and mère as a cousin relationship, not a direct historical relationship. For the purposes of this assignment, we are interested in English-French word pairs where the French word is the etymon of the English one; the more distant cognate relationship is not sufficient. (Of course, in general, English loanwords from French will have French cognates - unless the English word's French etymon became obsolete in French after the time of the borrowing.)

At first glance, it looks like we could extract words with French etyma by limiting our search to entries with "French" in the etymology field. However, this would yield the union of words with French etyma and words with French cognates. In order to restrict our results to English words with French etyma, we need to take into account the conventions that the OED editors use to indicate borrowings, which they term adaptations. Here are the ones that we've come across in class; there may be others.

cf. confer (Latin for 'compare') - Best avoided, as very likely to give you cognates rather than etyma!
< This symbol for 'derived from' from historical linguistics seems to be disregarded in searches (contrary to what we said in class).
a., ad. adapted (from)
f. from

In addition to the above conventions, it is useful to know the labels for (the dialects of) French that are of interest to us.

Language Abbreviations Comments
Anglo-French AF.
Anglo Fr.
Anglo-Norman Anglo-Norm.
French F.
Law French Law-Fr.
Turns out to be irrelevant, since searches for "Law French" never yield etymological information
Middle French MF.
Old French OF
Old Northern French ONF.
done - part of ling300-a9-santorini.xls

Who does what

As we saw in class, searching the OED for large datasets is mind-bogglingly slow, for reasons that escape me. In order to make the work less frustrating, I've decided to back off from having you search for the loanwords. Instead, I've conducted the searches myself, using the language labels above, and I've saved the results of the searches to .pdf files (see below). You should be able to click on the links to the .pdf files, and either download them or view them directly in your browser. Each of you will be responsible for entering part of the data in each of the .pdf files into a spreadsheet (which I'll describe in more detail directly). The table below contains the details of which data you're responsible for; the numbers indicate the running item numbers in the first column in the .pdf files.

Student a9-AFr.pdf a9-Anglo.pdf a9-F.pdf a9-Fr.pdf a9-MFr.pdf a9-OFr-0001-1000.pdf
Apple 1-39 1-3 done TBA done 1-100 done
Benton 40-79 4-6 301-400
Bumgardner 80-119 7-9 601-700
Ecay 120-159 10-12 901-1000
Flynn 160-199 13-16 (sic) 1201-1300
Gammill 200-239 17-19 1501-1600
Greiner 240-279 20-22 1801-1900
Kramer 280-319 23-25 2101-2200
Lam 320-359 26-28 2401-2500
Leung 360-399 29-31 2701-2800
Levine 400-439 32-34 3001-3100
Mannion 440-479 35-37 3301-3400
Phuong 480-519 38-40 3601-3700
Song 520-559 41-43 3901-4000
Uckun 560-599 44-46 4201-4300
Weng 600-639 47-49 4501-4600
Staff 640-end (done) 50-end (done) 4801-4900

In order to record your data, please download the spreadsheet template for this assignment and enter the data into your copy. If you've already downloaded the spreadsheet, please re-download it. Since the .pdf files save you the time you would have spent formulating and conducting searches, I've added three columns to the spreadsheet for you to record phonological information. For clarity, the name of the revised spreadsheet contains a date - a9-template-1030.xls; the older copy didn't. The columns in the spreadsheet are explained in detail below, after some general instructions.

Please read the instructions carefully before starting on the assignment, as the quality of your submission will affect the quality of the database and that of the results that can be obtained from it.

Instructions for recording data

General instructions

Explanation of column headings

Student Your last name
Entry The OED entry (that is, the English word that was borrowed from French).
Date Record the date of first attestation as it appears in the OED (with any question marks or other notes, such as ante or circa). Instead of listing ordinary dates, the OED sometimes lists periods of English (such as eOE 'early Old English', OE 'Old English', lOE 'late Old English'). Just note them as indicated. We'll convert them to suitable numbers later on. It's easiest and most reliable to get this information from the .pdf files. If, for some reason, you are working with headword entries, be aware that the date of first attestation is not always the first date that you see.
Etymon What is the French source of the loanword? Copy (best to cut and paste if easy to do so) the relevant information, including the "a.", "ad.", or "f.". It's fine to give just the first of several orthographic variants and to disregard morpheme boundaries in the etymon. For some examples, compare the data in ling300-a9-santorini.xls with the corresponding .pdf files.
# Syllables The information for the next three columns will come from the entries for the words themselves, not from the .pdf files. As mentioned above, please include all words that aren't clearly monosyllabic. Sometimes, it is not clear exactly how many syllables a word has. For instance, do schwa or syllabic sonorants count as syllables or not? That is, does gardener count has having two or three syllables? Does castle count as having one or two syllables? The most useful thing to do is to follow the OED. If a glide is transcribed as a vowel (even one in parentheses), count it as a separate syllable. We will deal with glides and schwas in a uniform way once the data are collected and glides. If the OED isn't much help in a particular case, don't spend a lot of time on it. But do make an appropriate comment. See ling300-a9-santorini.xls for how I've dealt with some problematic cases.
Stress Record the stress for the first pronunciation given in the OED. Since we are dealing with British English, the stress may differ from your own. Please count stress from right to left. If the OED records no stress, but gives the stress recorded in the New English Dictionary (NED), use that. Else (if no stress is recorded at all), include the word, but record the stress as "n/a".
Variant stress For some words, the OED records two variant stresses (e.g. ma'dame, 'ma.dame). Use this column to record any such variant stresses, following the conventions of the previous column. Variant stress is not the same as secondary stress.
Comments Record any notes, questions, uncertainties, and so so regarding the word. See ling300-a9-santorini.xls for some examples.