Literary and linguistic computing: from authorship attribution to assessing language change

Literary and linguistic computing:
from authorship attribution
to assessing language change

Maciej Eder

Institute of Polish Language (Polish Academy of Sciences)

TU Digihum Talks, 18th November 2022

Overview

Stylistic fingerprint
Authorship attribution at a glance
Stylometry as a tool to assess language change

The notion of stylistic fingerprint

An area between language constrains and a freedom of choice
Undiscoverable with the naked eye
Beyond authorial control
Resistant to imitation, plagiarism and parody
Popular solution: frequencies of the Most Frequent Words

Words are not distributed evenly

It is a truth universally acknowledged, that a single man in possession of a good fortune, must be in want of a wife. However little known the feelings or views of such a man may be on his first entering a neighbourhood, this truth is so well fixed in the minds of the surrounding families, that he is considered the rightful property of some one or other of their daughters.

“My dear Mr. Bennet,” said his lady to him one day, “have you heard that Netherfield Park is let at last?” Mr. Bennet replied that he had not.

“But it is,” returned she; “for Mrs. Long has just been here, and she told me all about it.” Mr. Bennet made no answer.

“Do you not want to know who has taken it?” cried his wife impatiently. “You want to tell me, and I have no objection to hearing it.” This was invitation enough.

Function words = frequent words

It is a truth universally acknowledged, that a single man in possession of a good fortune, must be in want of a wife. However little known the feelings or views of such a man may be on his first entering a neighbourhood, this truth is so well fixed in the minds of the surrounding families, that he is considered the rightful property of some one or other of their daughters.

“My dear Mr. Bennet,” said his lady to him one day, “have you heard that Netherfield Park is let at last?” Mr. Bennet replied that he had not.

“But it is,” returned she; “for Mrs. Long has just been here, and she told me all about it.” Mr. Bennet made no answer.

“Do you not want to know who has taken it?” cried his wife impatiently. “You want to tell me, and I have no objection to hearing it.” This was invitation enough.

Word frequencies as language features

raw occurrences of words:

## 
##  the   to   of  and  her    i    a   in  was  she that   it  not  you 
## 4330 4161 3609 3584 2227 2065 1955 1880 1847 1710 1577 1532 1426 1357

relative frequencies:

## 
##  the   to   of  and  her    i    a   in  was  she that   it  not  you 
## 3.49 3.36 2.91 2.89 1.80 1.67 1.58 1.52 1.49 1.38 1.27 1.24 1.15 1.09

Rank/frequency dependence

Authorship attribution

Non-traditional authorship attribution

Given:

a text of uncertain or anonymous authorship, and
a comparison corpus of texts by known authors,

is it possible to find the ‘nearest neighbor’ among the available candidates?

Example: The Cuckoo’s Calling

A debut by an unknown author Robert Galbraith
Only 1500 copies sold
Supposedly written by J.K. Rowling

Stylometry: where it (usually) begins

##                       the  and   to   of    a  was    I   in
## coben_breaker        3.59 1.18 2.16 1.38 2.52 1.50 1.45 1.18
## coben_dropshot       3.59 1.18 2.12 1.27 2.38 1.57 1.50 1.04
## coben_fadeaway       3.93 1.45 2.20 1.21 2.31 1.32 1.33 1.20
## coben_falsemove      3.63 1.61 2.13 1.24 2.40 1.38 1.35 1.11
## coben_goneforgood    3.83 1.82 2.15 1.18 1.96 1.73 3.81 1.13
## coben_nosecondchance 4.10 1.59 2.27 1.21 1.99 1.76 3.85 1.15
## coben_tellnoone      4.10 1.79 2.03 1.25 2.18 1.42 3.50 1.16
## galbraith_cuckoos    4.52 2.27 2.49 2.18 2.14 1.66 1.13 1.38
## lewis_battle         5.05 3.40 2.14 2.14 1.96 1.51 0.90 1.28
## lewis_caspian        4.87 3.59 2.15 2.14 2.17 1.35 1.12 1.21
## lewis_chair          4.97 3.22 2.00 2.10 2.35 1.40 1.07 1.21
## lewis_horse          4.89 3.49 2.31 2.22 2.32 1.40 1.20 1.30
## lewis_lion           5.14 3.70 2.29 2.19 2.10 1.35 0.81 1.16
## lewis_nephew         4.48 2.86 2.07 2.23 2.31 1.57 1.18 1.35
## lewis_voyage         5.22 3.28 2.26 2.11 2.24 1.58 1.05 1.15
## rowling_casual       4.75 2.64 2.62 2.11 1.76 1.65 0.56 1.44
## rowling_chamber      4.41 2.34 2.35 1.88 2.00 1.48 0.88 1.17
## rowling_goblet       4.48 2.43 2.49 2.02 1.79 1.42 0.85 1.12
## rowling_hallows      4.70 2.47 2.24 1.87 1.45 1.13 0.53 1.00
## rowling_order        4.23 2.25 2.51 2.03 1.79 1.39 0.86 1.17

Table of computed distances

##                      coben_breaker coben_dropshot coben_fadeaway
## coben_breaker               0.0000         0.6176         0.6164
## coben_dropshot              0.6176         0.0000         0.6059
## coben_fadeaway              0.6164         0.6059         0.0000
## coben_falsemove             0.6225         0.6392         0.6006
## coben_goneforgood           0.7093         0.7192         0.6827
## coben_nosecondchance        0.7219         0.7220         0.6936
## coben_tellnoone             0.7002         0.7211         0.6798
## galbraith_cuckoos           0.8730         0.9298         0.8923
## lewis_battle                1.1151         1.0911         1.0984
## lewis_caspian               1.1005         1.0912         1.0829
## lewis_chair                 1.0981         1.0859         1.0729
## lewis_horse                 1.0711         1.0559         1.0531
## lewis_lion                  1.1282         1.1223         1.1275
## lewis_nephew                1.0982         1.1043         1.0877
## lewis_voyage                1.0683         1.0843         1.0618
## rowling_casual              0.9424         0.9849         0.9607
## rowling_chamber             1.0780         1.1059         1.0795
## rowling_goblet              1.0497         1.0624         1.0538
## rowling_hallows             1.0426         1.0496         1.0249
## rowling_order               1.0424         1.0661         1.0323

Hierarchical cluster analysis

Multidimensional scaling

Bootstrap consensus tree

Supervised machine-learning approach

## Testing a given candidate against imposters...
## 
## coben 
## 
## lewis 
## 
## rowling 
## 
## tolkien

##   coben   lewis rowling tolkien 
##    0.39    0.00    1.00    0.00

Stylometry beyond authorship attribution

Authorship profiling
Forensic linguistics
Register analysis
Genre recognition
Gender differences
Translatorial signal
Early vs. mature style
Style evolution
Detecting dementia
…

Aim of the study

Introducing a method to trace changes in language
No a priori assumptions as to the change-makers
Identifying the impactful words over time
Application on the Corpus of Historical American English (COHA)

Langauge change

There’s no clear definition of language change
A more feasible approach: tracing changes in language.
Any two text compared exhibit differences (e.g. in word frequencies), the same applies to any two samples of a language.
A provisional definition: a language change is an accumulative effect of significant changes in a language.

The dynamics
of change

The dynamics of language change

Purely linear trends – not very probable.
Rather, epochs of substantial stylistic drift follow periods of stagnation.
The goal: observing the accelleration of language change.

What is responsible for a change?

A relatively small number of words that reflect changes in societies?
- (e.g. horse > automobile > car)
A aggregation of numerous features of individually weak discrimination power?
- (e.g. the signal of authorship)

Data and methods

Corpus of American Historical English (COHA)
- ca. 400 million words
- different genres (fic, mag, news, nf)
- covering the years 1810-2009
A supervised classification
- Nearest Shrunken Centroids classifier
- 1,000 most frequent words

Why does the change periodically accelerate?

A naive assumption that the change can be linked to particular events in history and/or culture.
However, the impact of major historical breakthroughs on language cannot be ruled out.

Who’s responsible for the change?

Distinctive features

Distinctive features: the biggest variation

So, which words are they?

##  [1] "the"       "and"       "week"      "that"      "'s"        "last"     
##  [7] "is"        "be"        "of"        "it"        "we"        "i"        
## [13] "to"        "was"       "mr."       "our"       "my"        "been"     
## [19] "not"       "u.s."      "you"       "new"       "upon"      "there"    
## [25] "has"       "says"      "war"       "york"      "this"      "n't"      
## [31] "will"      "s"         "which"     "for"       "had"       "very"     
## [37] "have"      "said"      "are"       "in"        "city"      "made"     
## [43] "me"        "a"         "president" "her"       "us"        "she"      
## [49] "his"       "by"        "united"    "or"        "at"        "but"      
## [55] "q!"        "party"     "committee" "your"      "so"        "as"       
## [61] "street"    "today"     "men"       "would"     "with"      "about"    
## [67] "heart"     "do"        "any"       "'d"        "out"       "all"      
## [73] "if"        "great"     "army"      "up"

So, which words made history?

Words cannot simply “disappear”

Personal pronouns (“social” words)

Conclusions

The language change seems to be a fluctuation of rapid and slow changes.
Numerous features responsible for the change, rather than few strong predictors.
- unexpectedly many pronouns among the predictors
- increasing power of “social” function words (personal pronouns)
Grand question: do we observe a language change, or a change in style? (e.g. a general taste of an epoch)