A paragraph can be a list of lists of lists

Last week, a pair was a list with two elements and a triple had three. Nested lists can also contain inner lists of different lengths. A paragraph can contain sentence lists; a sentence can contain word lists; a word can contain character lists. Each level answers a different question.

Paragraph = [Sentence1, Sentence2, ...]
Sentence1 = [explode("It"), explode("was"),
             explode("Monday"), explode("morning.")]
# first(Paragraph) is a sentence list
# first(first(Paragraph)) is a character list for its first word
# first(first(first(Paragraph))) is its first character

These expressions depend on the representation. Our original dataset stores word RECORDS with a Word field, not already-exploded character lists. For that version, get the first sentence’s first word record, read its Word field, then explode it. A number such as a student ID cannot be unpacked with first as though it were a list.

The original paragraph has five sentence groups of 4, 7, 8, 21 and 25 word cards. We use terminal full stops in these supplied cards to build the groups once, then reuse the sentence lists. This is a dataset convention, not a complete sentence splitter for abbreviations, quotes or general English.