In 2015, there are many means for collecting dialect evidence that weren't available when the Linguistic Atlas Projects (LAP) began sending fieldworkers to survey native informants in American communities in the 1930s. For instance, we now have the ability to create surveys that online users can complete at their leisure, or we can repurpose language intended for communication between "friends" and followers on Twitter and other social media to study dialectal variation. These methods mitigate some of the drawbacks of LAP methods, for instance, expenses associated with travel; the difficulty, at times, of finding informants willing and able to submit to long interviews with a stranger; and the time and cost of transcription. In doing so, these newer methods often provide a much quicker route to collecting, analyzing, and publishing dialectal data.

In spite of these advantages, there are some still benefits to be had by collecting data the "old-fashioned way", that is, by going out and talking to people, as evidenced by the most recent incarnation of the Linguistic Atlas tradition, namely, the Linguistic Atlas of the Western States (LAWS). First, unlike online surveys, which often have preselected responses from which users choose (much like a multiple choice test), LAWS interviewees are free to answer in any way they see fit, and, in doing so, sometimes offer linguistic variants that we have no knowledge of and that many of us in the area of dialectology seek. At the same time, LAWS fieldworkers do rely to a great extent on worksheets created before the onset of fieldwork, which not only guide the interview process but provide some clue to semantic relations held by responses that might not be readily transparent if they were removed from this context. Additionally, because many of the questions that LAWS fieldworkers ask pertain to objects with a physical presence in the homes of their informants, the relationship between "words and things" can be clearer in these interviews than when they are collected using other methods. Finally, it is often the case in what amount to the intersections of two or more very different lives in small windows of time and space that interviews do go off course for one reason or another: there are interruptions from the outside world, an informant loses patience with the line of questioning and begins going off on tangents, a fieldworker decides to try something new to gather enough momentum to make it to the finish line of a long interview, etc. And sometimes, these deviations occur similarly enough times in enough interviews that researchers can actually salvage something from these scrap heaps to support earlier ideas about the distribution of familiar linguistic variants or to shed light on a facet of language use previously unknown to them. Or, in the best of cases, both. 

To illustrate, this post investigates lexical variants associated with food preservation that were culled from a subcomponent of LAWS called the Linguistic Atlas of the Middle Rockies (LAMR). Unlike my previous blog posts on the language of LAMR, which have focused on explicit targets of the LAWS worksheets (namely aspen, gutters, and moonshine), there is no prompt in the worksheets concerning the act of preserving food; instead, this post relies on "collateral data" that emerged at different points in the interviews, in particular, during questions about types of crops that are or were important to ranch life in rural communities in Colorado, Utah, and Wyoming.*

Not surprisingly, at least to someone who grew up referring to the process of storing food in a jar as "canning" and only relatively recently bothered to wonder why, LAMR informants tended to use the word "can" in reference to food preservation. And, as illustrated by the following excerpts from the corpus, they used several variants of the word:

1. "In the earlier days, they used to can it because they didn't have any freezers or anything to keep it. They'd cut the meat up and, and take the bone out you know and can the meat." Lake City, CO
2. "Oh yes, my mother canned everything." Grantsville, UT
3. "There was a cellar for vegetables and canned foods but it wasn't under the house." Rock River, WY
4. "...and they're very hard to get ready for canning." Rock River, WY
5. "I don't remember mom a-canning pork but I imagine she did." Clearmont, WY

Of the 56 informants in the LAMR corpus who talked about preserving food, 47 used some form of the word "can" to do so: 34 informants used the verb forms can, cans, canned, canning (as in #1, #2, and including the a-prefixed form a-canning in #5) a total of 88 times; 33 used the adjectival forms canned (as in #3) and canning a total of 53 times; 8 used the gerund form canning (as in #4) a total of 9 times; and 7 used the noun forms can and cans a total of 10 times (although never clearly in reference to glass jars).

Speaking of jars, LAMR has a relatively straightforward distribution of jar and jars used in the food preservation context in which 10 people use the forms 17 times and always as a noun. Variants of "put up", which include put up and putting up, and may include an intervening element, as in "...then just fruit and they put them up fresh you know" (Duchesne, UT).

One set of lexical variants to emerge from collateral data on canning in LAMR consists of lemma of "bottle," which were used by 12 informants a total of 30 times in reference to food preservation, as illustrated by the following excerpts:

1. "You'd kill them and then you'd bottle the meat." (Ephraim, UT)
2. "...most of the fruit we had they bottled." (Manti, UT)
3. "...and they did an awful lot of bottling fruit and vegetables." (Ephraim, UT)
4. "Then she could just open the bottle of sausage and heat it up again." (Hoytville, UT)
5. "That was their main, their main diet is canned or bottled fruit...We never called it canned." (Lyman, UT)

The term occurs most often as the verb forms bottle, bottled, and bottling (as in #1, #2, and #3, respectively), being used by 10 informants a total of 18 times. It also occurs as the noun forms bottle (as in #4) and bottles, being used a total of 9 times by 5 informants. Finally, it appears as the adjectival form bottled (as in #5) a total of 3 times by 2 informants. Thus, fewer people use a form of "bottle" than use a form of "can", and variants of "bottle" are also more restricted with respect to the grammatical role they fill in LAMR compared to "can." Finally, as figure 4 shows, "bottle" is much more constrained in terms of geographical expanse.   

While more data would be necessary before making a strong case for the rather limited geographical distribution of "bottle" that is merely suggested by this figure, such results show the role that collateral data can have in pointing us to new areas of inquiry and the value of continuing to talk to the people.

-----------------------------------------------------------------------------------

*As part of this analysis, I looked only at places in the interviews where informants were discussing foods that had been preserved and ignored discussions of storing liquids, e.g. milk, water, and cooking oil.