THE Research

Using a rigorous and replicable analytic method, we offer a consistent analysis of the President’s public discourse on immigrants and immigration. On the basis of over 300 speeches and 5000 tweets, we located a single narrative that he articulates when speaking or tweeting about immigrants. His plot is the classic "America as Fortress" trope. Trump asserts that Fortress America is under attack and that its cities and towns have been overrun by ruthless invaders. Trump characterizes Mexico as the enemy that sent unauthorized immigrants as invaders. Trump represents himself as the hero, Hillary Clinton as one of the corrupt and sniveling politicians that led to this state of affairs, and US Latinos—namely US citizens—as the immigrants’ progeny that he disdains.

In this preliminary[1] report we exemplify the major metaphors Trump uses to articulate his narrative (see FINDINGS). We list the metonyms (code-words) that he uses to make his narrative forceful and misleading: he associates MS-13, the notorious gang, with all Latino gangs and even all young Latinos who wear certain clothing as a urban fashion statement. He extends the term criminal alien, namely immigrants who commit felonious crimes, to all unauthorized immigrants. He also refers by name to about twelve Americans who were the victims of major crimes, as representative of the imminent danger to all Americans and their family that immigrants pose. Finally, he expresses disdain toward US-born children of Latino immigrants, saying their increase should be halted, that they are not “our children,” and that they are not entitled to US citizenship.


Critical Discourse Analysis based on Conceptual Metaphor Theory

This study of President Trump’s discourse did not start with a hypothesis, but with the premise that his articulated expressions of conceptual metaphors can be located and described in a corpus of texts. The method is inductive, not deductive. The discourse unit that governs commonly used patterns of inference is metaphor. It is part of the conceptual system shared in large part by speakers of the English language and encoded in part in the ways Americans speak. When compiled in a coherent order, these metaphors can reveal an internally self-consistent constellation of concepts that make up the perspective the president articulates when speaking or tweeting.

As a speaker of American English, the president spoke and tweeted using shared conceptual metaphors to build his message on the basis of commonly shared understandings. Our empirical discourse analysis locates and interprets the conceptual metaphors that appear in a text. A key challenge in this scholarship is how to justify one and not another interpretation of a body of material. How can scholars avoid the biases to which discourse analysis on politicized topics is susceptible? The skeptic will rightly dismiss any analysis that exhibits selection bias, that is if it appears that the investigator has ‘cherry-picked’ data, choosing only the discourse material that suits his or her purpose. Second, the skeptic will reject any analysis that appears to have an interpretive bias. This occurs when the analyst builds an interpretation around a striking metaphor that is not representative of the corpus. To forestall these biases, a strict protocol was used to analyze meaning-laden discourse.

To avoid selection bias, the research team used corpuses that other scholars developed and compiled; future scholars can also retrieve these sources. *

We drew Trump's 2015 campaign speeches from:

We drew his 2016 campaign speeches from 

We drew Trump' presidential speeches from:

For Trump’s tweets, we relied heavily on the methods laid out in:

To avoid cherry-picking, the team systematically and comprehensively identified and coded all of the many hundreds of instances, ‘tokens,’ of the conceptual metaphors occurring in the president’s discourse. Every metaphor relevant to the topic in the corpus (i.e. involving immigration, the State, all pertinent actors, actions and processes, and institutions) was coded.

To avoid interpretation bias, a team, not an individual, studied the publicly articulated vernacular discourse, with the aim to seek the common denominator public interpretation of metaphoric material. A careful protocol is followed.[2] After the full set of data were coded, then the team works to develop an accounting of all the semantic relationships expressed among the hundreds of tokens. The goal is to locate the small set of metaphors employing greatest number of the tokens that coherently and consistently captures our reading of the president’s public discourse.

Conceptual metaphors on a given topic tend to form patterns that analysts can present in various ways. We chose a narrative summary. A narrative is a succinct integration of the conceptual metaphors, or the building blocks, of the president’s articulated worldview. The narrative composed is the team’s best effort to capture the president’s worldview when he publicly spoke or tweeted about immigrants and immigration. This protocol lends itself to replication by other researchers, and is consistent with the methods of other corpus metaphor analysts (e.g. Steen, et al. 2010).




Santa Ana, O. 2002. Brown Tide Rising: Metaphoric Representations of Latinos in Contemporary Public Discourse. Austin, TX: University of Texas Press.

Santa Ana, O, S.L. Treviño, M. Bailey, K. Bodossian, and A. de Necochea. 2007. A May to Remember: Adversarial images of immigrants in U.S. newspapers during the 2006 policy debate. Du Bois Review: Social Science Research on Race, 4.1, 207­–232.

Santa Ana, O. 2013. Juan in a Hundred: Representation of Latinos on Network News. Austin, TX: University of Texas Press.

Santa Ana, O, K, H. Waikuweit and M. E. Hernandez. 2017. Blood, soil and tears: Conceptual metaphor-based critical discourse analysis of the legal debate on US citizenship. Journal of Language and Politics, 16.2, 149–175.

Steen, G.J., Dorst, A.G., Herrmann, J.B., Kaal, A., Krennmayr, T., & Pasma, T. (2010). A method for linguistic metaphor identification: From MIP to MIPVU. Amsterdam; Philadelphia: John Benjamins.


[1] These preliminary finding are so robust, that we do not foresee major changes in the final analysis. The final report will be completed in January 2018.

[2] Santa Ana 2002, chapter 2; 2013, pp. 26­–29, chapters 6 and 7; Santa Ana et al. 2007, pp. 4–5; 2010, pp. 87–93; 2017, pp. 7–10.