9 March 2019. Puzzling out a word jumble, I’m writing a python script to search a grid for words. Step one is to compile a list of legal bigrams in English. Bigrams are two letters that go side-by-side. So the letter <Q> in English has a limited list of bigrams. We see <QU> as in quit, <QA> as in Qatar (and a few others if you allow very rare words).

I found a huge list online of English words compiled from web pages. 2.5 megs of text file! Here is the resulting python dict of bigrams:

{'A':['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z'],
'B':['A', 'B', 'C', 'D', 'E', 'F', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'R', 'S', 'T', 'U', 'V', 'W', 'Y', 'G', 'P', 'Z', 'Q'],
'C':['A', 'I', 'K', 'T', 'U', 'E', 'O', 'Y', 'H', 'C', 'L', 'M', 'N', 'Q', 'R', 'S', 'D', 'B', 'W', 'Z', 'G', 'P', 'F'],
'D':['V', 'W', 'E', 'I', 'O', 'L', 'N', 'A', 'U', 'G', 'Y', 'R', 'P', 'C', 'D', 'F', 'H', 'J', 'M', 'S', 'T', 'Z', 'B', 'K', 'Q'],
'E':['H', 'R', 'D', 'N', 'E', 'S', 'M', 'Y', 'V', 'L', 'A', 'C', 'I', 'P', 'T', 'K', 'Z', 'U', 'G', 'W', 'B', 'F', 'O', 'X', 'Q', 'J'],
'F':['F', 'T', 'A', 'U', 'O', 'E', 'I', 'Y', 'L', 'G', 'R', 'S', 'W', 'Z', 'N', 'V', 'H', 'B', 'K', 'D', 'M', 'J', 'P', 'C'],
'G':['I', 'E', 'H', 'L', 'N', 'A', 'Y', 'O', 'R', 'M', 'U', 'S', 'D', 'G', 'K', 'P', 'B', 'W', 'T', 'F', 'C', 'V', 'J', 'Z'],
'H':['R', 'E', 'L', 'M', 'I', 'Y', 'O', 'U', 'A', 'T', 'N', 'S', 'W', 'B', 'P', 'Z', 'G', 'C', 'F', 'D', 'H', 'J', 'K', 'V', 'Q'],
'I':['C', 'T', 'N', 'S', 'O', 'E', 'A', 'Z', 'R', 'L', 'D', 'U', 'P', 'G', 'B', 'V', 'F', 'M', 'I', 'X', 'K', 'Y', 'W', 'H', 'Q', 'J'],
'J':['E', 'O', 'U', 'A', 'I', 'H', 'J', 'R', 'Y', 'P', 'D', 'M', 'W', 'L', 'T', 'N', 'B', 'K'],
'K':['A', 'H', 'E', 'I', 'Z', 'M', 'N', 'B', 'S', 'L', 'O', 'C', 'K', 'P', 'R', 'T', 'U', 'W', 'Y', 'D', 'F', 'G', 'J', 'V'],
'L':['F', 'L', 'U', 'I', 'O', 'E', 'Y', 'A', 'M', 'T', 'S', 'N', 'V', 'C', 'D', 'B', 'G', 'H', 'P', 'R', 'K', 'W', 'J', 'Q', 'Z', 'X'],
'M':['A', 'P', 'E', 'B', 'I', 'O', 'H', 'U', 'Y', 'M', 'S', 'T', 'F', 'L', 'W', 'N', 'R', 'C', 'G', 'V', 'K', 'D', 'J', 'Z', 'Q'],
'N':['I', 'A', 'C', 'E', 'D', 'T', 'U', 'O', 'S', 'R', 'G', 'Y', 'M', 'N', 'Z', 'L', 'P', 'K', 'F', 'H', 'Q', 'B', 'J', 'V', 'X', 'W', '-'],
'O':['L', 'N', 'R', 'S', 'I', 'M', 'T', 'U', 'G', 'O', 'W', 'A', 'B', 'D', 'H', 'V', 'X', 'C', 'K', 'Z', 'P', 'Y', 'E', 'F', 'Q', 'J'],
'P':['E', 'T', 'O', 'Y', 'I', 'H', 'S', 'R', 'A', 'N', 'U', 'L', 'P', 'M', 'J', 'B', 'D', 'F', 'W', 'K', 'C', 'G', 'V', 'Q'],
'Q':['U', 'I', 'A', 'R', 'E', 'O', 'Q'],
'R':['D', 'O', 'U', 'E', 'A', 'I', 'T', 'Y', 'R', 'S', 'V', 'M', 'B', 'P', 'G', 'N', 'H', 'L', 'F', 'W', 'C', 'K', 'J', 'Q', 'X', 'Z'],
'S':['C', 'T', 'A', 'E', 'S', 'I', 'G', 'H', 'K', 'O', 'M', 'U', 'F', 'Q', 'V', 'Y', 'P', 'L', 'N', 'B', 'W', 'D', 'R', 'J', 'Z'],
'T':['E', 'I', 'O', 'H', 'A', 'T', 'U', 'C', 'N', 'S', 'R', 'M', 'L', 'Y', 'B', 'P', 'F', 'W', 'K', 'Z', 'D', 'G', 'J', 'V', 'Q', 'X'],
'U':['A', 'S', 'L', 'R', 'C', 'M', 'N', 'D', 'T', 'E', 'V', 'P', 'Z', 'B', 'I', 'O', 'X', 'G', 'K', 'F', 'Y', 'W', 'J', 'H', 'Q', 'U'],
'V':['A', 'E', 'I', 'O', 'U', 'Y', 'S', 'R', 'C', 'L', 'V', 'N', 'Z', 'D', 'K', 'G'],
'W':['O', 'H', 'A', 'E', 'I', 'L', 'N', 'S', 'T', 'R', 'M', 'U', 'Y', 'B', 'P', 'W', 'D', 'F', 'K', 'C', 'G', 'Z', 'Q', 'V', 'J'],
'X':['I', 'A', 'Y', 'T', 'E', 'O', 'U', 'M', 'P', 'C', 'B', 'F', 'H', 'L', 'S', 'W', 'R', 'D', 'K', 'N', 'G', 'Q', 'Z', 'V'],
'Y':['S', 'M', 'A', 'R', 'C', 'P', 'G', 'I', 'L', 'N', 'D', 'T', 'X', 'O', 'E', 'Z', 'U', 'F', 'W', 'H', 'B', 'Y', 'K', 'V', 'J', 'Q'],
'Z':['E', 'A', 'U', 'Z', 'I', 'O', 'L', 'G', 'Y', 'R', 'H', 'T', 'N', 'B', 'D', 'P', 'K', 'C', 'M', 'V', 'S', 'F', 'W']

And here is the code to get the bigrams (my file of words is called web2.txt, and each word is on a separate line). In order to limit the bigrams to a list of unique letters, I use set().

import os

path = os.getcwd()
path += '/web2.txt'

bigrams = {'A':[], 'B':[], 'C':[], 'D':[], 'E':[], 'F':[], 'G':[], 'H':[], 'I':[], 'J':[],
           'K':[], 'L':[], 'M':[], 'N':[], 'O':[], 'P':[], 'Q':[], 'R':[], 'S':[], 'T':[],
           'U':[], 'V':[], 'W':[], 'X':[], 'Y':[], 'Z':[]}

with open(path, 'r') as allwords:
    words ='\n')

for letter in bigrams.keys():
    letter = letter.upper()

    for word in words:
        word = word.upper()
        if letter in word:
            if word.index(letter) < len(word):
                    nextletter = word[word.index(letter) + 1]
                    if nextletter not in set(bigrams[letter]):
                except IndexError:

    print('\'{0}\':{1}, '.format(letter, bigrams[letter]))


March 6. An interim step in making a semantic map of Old English is producing bigrams. Bigrams are pairs of words. In order to build a social network of words, you need to know which words connect to one another. For example, in Beowulf, the word wolcnum ‘clouds’ almost always sits next to under ‘under’.

By comparison, the epic poem Judith has no clouds in it. And the homilist Ælfric never uses the phrase under wolcnum.

Here is a screen shot of words that follow ic ‘I’ in the poem Beowulf. So, the first is “ic nah.”

You can see that there are 181 instances of ic, although only 80 are unique. In other words, some bigrams are repeated. The second word of the bigram is printed again in red, and passed to a part-of-speech tagger. The blue text is the tagger’s best guess, and it also returns the part-of-speech most cited by dictionaries. As I plan to discuss in an article, ic is very rarely followed by a verb.

We can discover a great deal about poetic style by looking very closely at the grammar of Old English poetry. The grammar is the unfolding in time of images and ideas and asides and so forth. Grammar describes how the words affect you in order as you read.