Using the text search
First, it’s important to note that you can only search the transcriptions
of poems where they have actually been transcribed. Only 2181 poems of
the 15548 in the collection have been transcribed (these are mostly
those published on the decade years—1820, 1830 and so on—but there are others scattered
across
the collection). For untranscribed poems, you can search the metadata, first
line, illustration descriptions etc., but not the poem transcription or the
information on poetics and style.
This is a stemming search engine so, generally speaking, if you search for a word
such as love, the
search engine will apply stemming and return related
forms such as loving and loves. For finer
control, there are two wild-card characters that can be used in searches:
asterisk (*) and question mark (?). An asterisk represents zero or more
characters; a question mark represents a single character. A wild-card
search allows you to truncate endings, so that a search for
usur* will return results that include
usury, usurie, and
usurer. The wild card can also be used within a
word to return all possible variations in that position. For example, a
search for w*ld would return
wild, world, and
withheld, and so on, while a search for w?ld
would return only four-letter words such as wild or
weld. Combining internal and terminal wild cards
would return more variants. For example, w?ld* would
yield results that include wild,
wildest, and wilderness.
You can also use plus and minus signs to specify that a term must or must not be in
the results.
For example, searching for +love +like -hate will find documents that
contain both love and like
but not hate.
Using the filters
The search filters enable you to generate or answer sophisticated research questions.
Here is a worked example:
Imagine that you would like to track the prevalence of the stanza rhyme-scheme abab
in periodicals over the Victorian period: in what decades is it most popular? (Of
course it is
important to remember that these dates are dates of publication, not of composition;
older poems
are frequently republished in these periodicals.)
We can gather some evidence on this
question by using the search filters.
Before we start, we need to acknowledge that only transcribed poems have information
about
their dominant rhyme-scheme, so the first thing to do is to check the Transcribed
checkbox in the Transcription status
field.
Now we can gather information for the decade 1820-1829:
- Put
1820
in the Publication / Date From field, and 1829
in the Date To.
- Click on
Search
to find out how many transcribed poems there are in that period.
- Note the number of documents found (197 at the time of writing, February 2022).
- Now constrain the count by clicking on
Poem features
, and
type abab
into the Dominant rhyme scheme
box. From
the drop-down list, select abab
. This creates a filter for that
specific rhyme-scheme and checks it for you.
- Now do the search again, and note the number of poems (18 at the time of writing).
That means that in this collection, 18 * 100 / 197 = 9.14% of poems use the rhyme-scheme
abab
.
- Now do the same calculations for each subsequent decade-year. To get the total
number of poems, just uncheck the
abab
control, then check it
again to activate the filter.
The table below shows the results at the time of writing.
| Decade |
Total Poems |
Poems with abab |
Percentage |
| 1820s |
197 |
18 |
9.14% |
| 1830s |
186 |
27 |
14.52% |
| 1840s |
293 |
56 |
19.11% |
| 1850s |
180 |
30 |
16.67% |
| 1860s |
375 |
58 |
15.47% |
| 1870s |
241 |
33 |
13.69% |
| 1880s |
205 |
29 |
14.15% |
| 1890s |
261 |
36 |
13.79% |
As you can see, we don’t learn much of interest from this particular
operation, except that there is a brief small peak in usage of abba
in poems published in the 1840s.
Martin Holmes, February 2022