About

 

Human societies and communities may be organized in various forms, but one thing we share is the privilege of language and interaction.

In the 1960s, J.L. Austin proposed the idea of performative utterances: utterances that are not about being true or false but have actual consequences in reality. For example, when a judge says, “I sentence you to death,” he is performing the sentencing action; when a bride says “I do,” she is making a commitment to her marriage.

John Searle continued to develop this theory and analyzed promise-making in detail (Searle 1969).

Searle defined promise-making as a subtype of commissive speech acts and is accomplished by a proposition that expresses the speaker’s intention to do something in the future that the hearer wishes. A set of conditions and principles must be fulfilled to practice a sincere promise, such as the purpose of the act, the relative positions of the speaker and the hearer, and the degree of the commitment undertaken.

Common linguistic devices for promises include verbs and nouns that directly express commitment (e.g. promise, commit, vow, etc.), modal verbs (e.g. must, should, will, etc), and semi-modals (e.g. be going to, want to, need to, etc.) that express the speaker’s attitude.

This project explores how American politicians use these devices in their promises. Particularly, I aim interested in two questions: (1) have the presidents’ promise expressions changed throughout U.S. history? (2) Do governors across the country have similar preferences of promise expressions?

Modals and modality system

Modality is a grammatical category closely associated with promises and commitment as it marks the speaker’s ability, necessity, permissions, and obligation in utterances. This project follows Palmer’s (2001) modal system with two main categories, propositional and event. Propositional modality is further classified into epistemic and evidential; event modalities include deontic modality and dynamic modality.

Epistemic modality indicates speakers’ stance on the factual status of the proposition, often marked by modal verbs in English such as maybe, might, must; 

Evidential modality indicates speaker’s evidence for the factual status, marked by expressions such as it is said, and I heard; 

Deontic modality expresses obligations or permissions from external sources, marked by modal verbs such as must and should, and semi-modals such as have to, ought to; 

Dynamic modality indicates the speaker’s willingness or ability that comes from internal concerns, marked by modal verbs, such as will, can, and shall.

Based on the categorizations by Palmer (2001) and Biber et al. (1999), these common promise-relevant modals are grouped into three types: volition (willingness), obligation, and ability/possibility (epistemic modals may and might, as well as would are less relevant to promises and are found no occurrence of commitment-expressing cases in the current dataset.)

These modals, along with verbs and nouns expressing commitment (such as commitment, assure, promise, etc.), will be searched as promise-expressions in the database. The frequency results will be analyzed and visualized to answer the two main questions in this project.

DATASET

The current database includes U.S. presidents’ inaugural speeches and the State of the States Addresses (governors’ new year speeches) in 2019. The governors’ speech data are downloaded from the  What America’s Governors Are Talking About project (see Github) by the group FiveThirtyEight. Inaugural speeches are downloaded from The American Presidency Project hosted by UCSB. 

METHODS

Corpus Linguistic programs Lancsbox developed by Lancaster University is used to analyze the normalized frequencies of the promise-expressions and to generate the concordances. Since many modals can express multiple modalities (e.g. must as an epistemic/possibility modal and a deontic/obligation modal), annotations to distinguish these functions are yet to be added. 

The concordance lines extracted are then further analyzed with the Mallet Topic Modeling Tool. Results of the analysis provide a glimpse of the content of promises made by different presidents and governors.

VISUALIZATION

Visualization is realized through Fourish.

Frequencies of different expressions are visualized with grids of column charts, stacked area charts, stacked bar charts, and maps. For the governors’ promise section, an interactive map of the U.S. is created to show the main topics of promises made by governors across the country and the frequencies of promise expressions identified in the current project.