Alan Liu, August 3, 2015

DOI: 10.5072/FK2FN18G5G

• Anecdotal Evidence from U.K. (and Canadian) Scholars: Consultation via email on July 30-31, 2015, with Professors Melissa Terras (University College, London) and Andrew Prescott (University of Glasgow) indicates that the term "humanities" is primarily an Americanism that came into increasing usage in the U.K. relatively late in the 20th century. "Arts" was the previous and still widely used term in the U.K. for the humanities in the sense of literature, history, philosophy, etc. (along with their academic fields). "Arts" thus has a fuzzy generality that distinguishes it from "art" in the singular, which refers more narrowly to fine art, art practice, etc. Prefixed with the definite article, the phrase "the arts" may be even more fuzzily general in its overlapping reference to both "arts" and "art." (Professor Stéfan Sinclair [McGill University] reports that usage in Canada likely overlaps with both the U.K. and the U.S.  However, WE1S has not yet further researched usage in Canada and other Commonwealth nations.) The recent rise in the U.K. in usage of "humanities" (sometimes in conjunction with "arts" in the general sense) may be owing both to American influence and to institutional and funding needs for differentiation between the humanities and fine arts--especially after the Arts & Humanities Research Council was established in 2005 as distinct from the Arts Council created in 1994.

• Evidence from one U.K. newspaper (The Guardian): The WE1S research team used The Guardian's API - new version to locate all articles mentioning "humanities," "arts," and "the arts" in The Guardian from 1994 (the first year digitized articles appear to be available) through 2014.  (We also searched for articles mentioning "liberal arts" for symmetry with our searches of U.S. newspapers, even though the latter phrase is known to be primarily an Americanism.)  The following are counts of articles mentioning "humanities", "arts", and "the arts" in The Guardian (by year). The low count prior to 1999 is likely an artifact of the scope or searchability of digitized texts in The Guardian during its early years of digitization:

• For a deeper look, WE1S used its Guardian collection workflow first to scrape the full text of all articles mentioning "humanities" and "liberal arts" for these years (equivalent to our scraping of U.S. newspapers). Then, as an experiment, we separately scraped all articles mentioning "arts" and (separately) "the arts" from one early year in the range (1999) and one late year (2014). Whereas there are usually about 200 articles per year mentioning "humanities" during the later years of The Guardian, the number of articles mentioning "arts" is much higher:
• For 1999, there were 1351 Guardian articles mentioning "arts", and 454 articles mentioning "the arts".
• For 2014, there were 4,587 Guardian articles mentioning "arts," and 1,043 articles mentioning "the arts".
• (There is of course overlap between articles mentioning "arts" and "the arts".)
• We then human-read about 100 articles in each of the article sets for "arts" and "the arts," observing the main contexts/themes. Our initial finding was that the phrase "the arts" is likely the best search term to add to our searches on "humanities" and "liberal arts." "The arts" allows us to pick up a majority of articles about arts in the narrower sense of fine, visual, practicing, performing arts, but also an important minority of articles about arts in the more general humanistic sense of literature, history, philosophy, classics, art history, etc. By contrast, searching on "arts" (without the definite article) produced an order of magnitude more results. The relevant results overlap with those from a search on "the arts." The irrelevant results include a larger proportion of articles about "art" in the narrower sense of fine art, theatrical art, musical art, media art, etc., and also a fairly high number of articles on such wholly unrelated topics as "martial arts," "black arts," etc.
• Finally, after our human-reading gave us a sense that "the arts" is probably the best search term to add, we text-analyzed our full-text scrapes of articles mentioning "the arts" in The Guardian for 1999 and 2014. We used the Antconc tool to examine the most frequent words in the article set ("wordlist"), the most frequent bigrams beginning with the word "arts", and some sample key-words-in-context (KWIC) concordance results. The word frequencies and bigrams show that the context of "the arts" is usually the fine, practicing, or performing arts; but some examples in the concordance KWIC results confirm our sense from human reading that a minority of usages occurs in a more general humanities sense.

• Corpus Linguistics Evidence: 
• A different view of general patterns of linguistic usage of "humanities" and "arts" can be gained through corpus linguistics. While WE1S is not a corpus linguistics project (our researchers are not experts in this regard), we used the British National Corpus (BYU-BNC) (100 million words; coverage: 1980s-1993) to gain context on our questions about usage.  We searched the BNC for "humanities", "arts", and "the arts"; and we concentrated on looking at the sections of the corpus drawn from newspapers and magazines (looking at collocates in those sources). The following are screenshots of partial results. (The BNC corpus is unfortunately not broken into chronological segments, and so does not allow us to trace changes in usage between the 1980s and 1993):

"humanities":	BNC - "humanities" usage in magazines (too few newspapers to include)		BNC - "humanities" - collocates
"arts":	BNC - "arts" - in magazines	BNC - "arts" in newspapers	BNC - "arts" - collocates
"the arts":	BNC - "the arts" - in magazines	BNC - "the arts" - in newspapers	BNC - "the arts" - collocates

• Supplementary or corroborative evidence can be found in the Hansard Corpus (British Parliamentary speeches) (1.6 billion words; coverage: 1803-2005). This corpus is segmented chronologically and shows change by decade. (Note that the very early uses of "humanities" in the 19th century tend to mean something like "humanity" or "human kindness"):

• Supplementary or corroborative evidence can also be found in the Google Books British English archive (via Mark Davies's BYU corpus site):

• Canadian usage can be compared in the  Strathy Corpus (Canada) (50 million words; coverage: 1970s-2000s):

• Initial reflection on results from the BNC and other linguistic corpora indicates to us that:
• There has indeed been a rise in usage of "humanities" in the U.K. later in the 20th century;
• The "arts" terms primarily refer to the context of fine arts, etc.
• But there are usages of "arts" that overlap with the more general sense of the humanities.
• Meanwhile, on the other side of the Atlantic, the Corpus of Historical American English (COHA) (400 million words; coverage: 1810-2009) shows the following usage rise in "humanities" by decade. The spike beginning in the 1960s may corroborate the argument of Geoffrey Galt Harpham's The Humanities and the Dream of America (2011) that the humanities rose into prominence in the U.S. at the policy, ideological, and institutional levels as part of the Cold War era need to distinguish the American versus Soviet ideal.

• Conclusion:
• This preliminary study of linguistic usage indicates that:
• "humanities" arose in prominence as a general term in public discourse first in the U.S. from the 1960s on,
• "humanities" also rose in frequency later in the 20th century in public discourse in the U.K.
• "arts" in the U.K. refers primarily to the narrower domain of fine and practicing arts, but does indeed have a fuzzy overlap with the general sense of the humanities.
• (very tentative hypothesis:) usage of "arts" in the U.K. may be losing some of its generality as the term is increasingly positioned in conjunction with "humanities" (as in such compound phrases as "arts and humanities") so as to differentiate it.
• Practical conclusion: WE1S will search/scrape the phrase "the arts" for U.K. newspapers and magazines in addition to "humanities" (and "liberal arts"). Scraping "arts" (without the definite article) is too expensive in terms of the research time needed to process a much larger number of (often irrelevant) articles; and any gains over searching instead on "the arts" (which returns a smaller number of more relevant articles) appears to be vanishingly small. We are confident that searching on "the arts" (especially in combination with "liberal arts") will pick up most of the minority of arts articles that refer to the humanities. Topic modeling afterwards should help further disambiguate various senses of "arts" in the found material.
• We have no conclusion yet on linguistic usage in public discourse on the humanities and arts in the Canadian, Australian, New Zealand, or Indian press (pending further research). (Currently, WE1S is working on English-language public discourse. While 4Humanities has an interest in "global humanities," the theme of one of its past events, WE1S at present does not have the means or methodology [e.g., cross-linguistic topic modeling] to address public discourse in other languages.)

Additional Question: For Symmetry with Adding "The Arts" to Searches in U.K. Sources, Should WE1S Also Search on "The Arts" in U.S. newspapers and magazines (as well as "humanities" and "liberal arts")?

• Experiment: Alan conducted the experiment of searching/scraping articles mentioning "arts" and "the arts" in the New York Times for 1996.
• There were 3,116 articles mentioning "arts" (compared to 196 in the same year mentioning "humanities" and 111 mentioning "liberal arts"). 301 of these "arts" articles were duplicates of articles picked up by searching for "humanities" and "liberal arts".
• There were 3,112 articles mentioning "the arts."
• Human reading of a sample of the non-duplicate "arts" articles failed to detect any exceptions to the apparent rule that "arts" in the U.S. press refers to fine arts, performing arts, media arts, etc. as differentiated from "humanities."
• Practical Conclusion: For the purposes of the WE1S study of public discourse on the humanities, searching on "arts" in U.S. publications would not be effective. While it would widen the collection in a way that could prepare for future comparison of the arts and humanities, it would cost too much of the remaining WE1S grant to add a scrape on "arts" to the work on each U.S. publication. (E.g., adding a search and scrape on "arts" for the NY Times would likely cost about $1,000 worth of research assistant time.)