'"Everything on Paper Will Be Used Against Me": Quantifying Kissinger'

'"Everything on Paper Will Be Used Against Me": Quantifying Kissinger'

Kaufman
Micki

CUNY Graduate Center, United States of America
mkaufman@gc.cuny.edu

2014-12-19T13:50:00Z

Paul Arthur, University of Western Sidney

Locked Bag 1797
Penrith NSW 2751
Australia
Paul Arthur

Converted from a Word document

DHConvalidator

Paper

Long Paper

diplomatic history
sentiment analysis
topic modeling
visualization
network analysis

archives
repositories
sustainability and preservation
corpora and corpus activities
databases & dbms
historical studies
digital humanities - nature and significance
metadata
project design
organization
management
digitisation
resource creation
and discovery
semantic analysis
software design and development
text analysis
content analysis
interdisciplinary collaboration
linguistics
visualisation
spatio-temporal modeling
analysis and visualisation
data mining / text mining
English
rhetorical studies

Fixed in the public consciousness by countless appearances and interviews, and intimately associated with a wide range of affairs of state during the 1970s, including the US invasion of Cambodia, withdrawal from Vietnam, ‘opening to China’, détente with the Soviet Union, and the Watergate affair, the public persona of former national security advisor (1969–1975) and US secretary of state (1973–1977) Henry A. Kissinger has consistently fascinated historians.
1 A decidedly emotional champion of dispassionate ‘realpolitik’, a Cold Warrior and ‘secret swinger’,
2 questions about Kissinger’s fascinating paradoxes and internal contradictions have been a significant focus of the study of ‘Kissingerology’ (Hanhimaki 2003).

Trying to understand what at first glance can appear to be apparently incompatible motives and behavior, historians soon encounter a second problem—one of scale. A classic ‘big data’ catch-22, the extensive and vast array of documents, diplomatic cables, transcripts, and other correspondence available for study greatly complicates the task of historically situating and interpreting Kissinger. This deluge of information is an increasingly common frustration for historians of the 20th century, and as larger and larger archives of human cultural output are accumulated, scholars are beginning to adapt, develop, and employ tools, methods, and interpretive frameworks from fields like computational linguistics, visual design, and textual studies that can overcome ‘information overload’ and facilitate new historical interpretations of ‘big data’ archival collections.

A topic model of the DNSA Kissinger ‘memcons’. In this force-directed diagram, the documents are distributed according to their weight in the topic model and colored according to their former classification status. Formerly ‘Top Secret’ documents are colored blue, ‘Secret’ documents are colored yellow, and ‘Confidential’ are colored magenta. The size of the circles represents the number of participants.

As detailed on the website http://www.quantifyingkissinger.com, this project is an historical interpretation of the DNSA (Digital National Security Archive)’s Kissinger Collection, using techniques like data visualization, topic modeling, sentiment analysis, and word frequency and collocation analysis to facilitate a mix of ‘distant reading,’ ‘networked reading’, and ‘close reading’. The DNSA’s Kissinger Collection is a small but substantial subset of the vast volume of material generated by the Nixon administration, a collection comprising over 18,000 declassified meeting memoranda (‘memcons’) and teleconference transcripts (‘telcons’) detailing Kissinger's correspondence during the period 1969–1977, including a wide range of extremely useful archivist-supplied metadata (e.g., date, classification status, origin, etc.) that facilitates more detailed and nuanced explorations into the text analytics.
Faced with the seemingly impossible task of truly understanding a man who appears to embody a host of mutually exclusive dichotomies, historians who grapple with some of the more paradoxical aspects and events of Kissinger’s policies and persona often seek to explain inconsistencies and/or emotionality as either a logical failure of his political philosophy or as irrational divergence from an otherwise ubiquitous, calculating rationality. Countering the decentralized, ‘accidental’ view of emotionalism in most of the historiography, historian Barbara Keys’ ‘The Emotional Statesman’ places Kissinger’s emotionality in the center of her analysis, arguing that it was often the rhetorical and political flexibility emotionality offered that led Kissinger to such behavior.

Figure 1. Scatter plot graph: Frequency of words found in colocation with the word ‘laughter’ in the Kissinger memcons, in order of frequency (x-axis), by colocation (Mutual Information) score (y-axis).

Exploring Kissinger’s ‘predispositions’ towards a Soviet-preponderant worldview as a ‘natural’ emotional reaction resulting from his intense personal relationship with Soviet ambassador Anatoly Dobrynin (as detailed by Keys), the project’s word collocation analysis of the phrase ‘[laughter]’
3 surfaces evidence of the unique and complex nature of their friendship. To Keys, this relationship was a compelling emotional factor in Kissinger’s seeming unwillingness to view the Vietnam War as anything other than a Cold War conflict, one that nevertheless served an emotional purpose: reinforcing a sense that matters remained in his and Dobrynin’s capable hands to resolve no matter how arduous and unpleasant the peace talks with North Vietnamese envoy Le Duc Tho became.

Figure 2. Force-directed diagram: ‘Bombing’ word correlation/collocation in telcons (yellow) and memcons (blue). Words colorized by hand relate to Vietnam (green) and other countries in Indochina (red).

In addition to analyzing the impact of Kissinger’s social relationships on his geopolitical outlook and outcomes, the project examines also his attitudes towards matters of secrecy and violence. A number of word collocation analyses reveal patterns suggestive of a highly selective, event-driven use of the ‘private’ telcons (which Kissinger expected to remain within his possession after his departure from the White House) versus a more diffuse, persistent use of the official, redacted memcons—as, for example, when both the word ‘bombing’ and one of a number of names for various countries in Indochina were found.

Figure 3. Line/scatter plot graph: Frequency of ‘[Cc]ambodia’ and ‘bombing’ word collocation with collocation distance) on a timeline.

The initial text analysis data averages word frequencies irrespective of the timeline, and more detailed word frequency analysis along a timeline facilitated a ‘networked reading’ approach
4 in which documents are identified for closer reading and subsequent analysis through the text analysis data, and related according to commonalities in date, word frequency, topic weight, and/or subject and participant metadata.

Figure 4. A selected ‘network’ of a group of conversations and memos from 1970 in which ‘[Cc]ambodia’ and ‘bombing’ are found in collocation, connected by participant. Edges are labeled with the date of each individual’s participation.

A complete historical understanding must by necessity include this sort of close reading of relevant documents, but a few interesting bits of possible evidence from the project’s distant and networked reading already call for further analysis: hints of ‘triangular diplomacy’ written in the proximity of graphed nodes relating to negotiations over Indochina with the Soviet Union and China, and clustering of various ‘secrecy’ topics highly suggestive of a classification scheme based upon both national security priorities and/or those of Henry Kissinger himself. Topic clusters in the model represent SALT Negotiations between the USSR and the US, Nixon’s journey to China, the Paris Peace Talks to end the war in Vietnam, the OPEC oil embargo, crises in the Mideast and shuttle diplomacy, the coup in Greece and subsequent invasion of Cyprus by Turkey, and the Rhodesian Bush War.

Figure 5. A section of the topic model visualization of the ‘memcons’. In this force-directed diagram, the documents are distributed according to their weight in the topic model and colored according to their former classification status. Formerly ‘Top Secret’ documents are colored blue, ‘Secret’ documents are colored yellow, and ‘Confidential’ are colored magenta.

The examinations of this project have not been limited to historiographical inquiries of Kissinger’s foreign policy, nor to his personal/emotional motivations. By combining computational and emotional history approaches, insights about the man and the geopolitical focus of the administration he served emerge—new avenues for understanding themes of seduction, secrecy, humor, and violence. Beyond detailing and exploring existing field conversations about former US secretary of state Henry Kissinger, the project has begun to surface deeper understandings and new questions—about how this new kind of distant knowledge is formed, and the possible ramifications for historical research in general. This application of computational techniques to the study of 20th-century US diplomatic history has generated useful finding aids for researchers, provided essential testing grounds for new historical methodologies, provoked new questions, and prompted new interpretations of Kissinger and the individuals with whom he corresponded.
Notes
1. Kissinger was the only official ever to hold both offices simultaneously, from 22 September 1973 to 3 November 1975.
2.
LIFE magazine, 28 January 1972
, https://books.google.com/books?id=D0AEAAAAMBAJ&pg=PA70.

3. As well as related (but different) search and wildcard phrases like ‘[laughs]’, ‘[laugh*],’ etc.
4. An approach described by Dr. Lisa Marie Rhody.

Bibliography
Hanhimaki, J. (2003). ‘Dr. Kissinger’ or ‘Mr. Henry’? Kissingerology, Thirty Years and Counting*.
Diplomatic History,
27(5) (November).