In Search of the Benefits of Haag & Stern’s “In Search of the Benefits of Learning Latin”

2003 saw the publication of “In Search of the Benefits of Learning Latin” by Ludwig Haag and Elsbeth Stern in The Journal of Educational Psychology. Frothing opponents of Latin studies seized upon the study (beware the logical argument depending on one study), and even proponents remain concerned by it.

Sixteen years later, Latin studies are holding strong in the United States, and European authorities have recognized that Latin and Ancient Greek are “compulsory subjects or options for those students attending the most academic types of schools.” It sounds like Haag and Stern have faded into obscurity, though Google Scholar notes 60+ citations.

It needs to be said: their publication has much dubious content. If you encounter opponents of Latin referencing it, keep these ten things in mind:

1. They overgeneralize the implications of their findings right off the bat (abstracts have power).

Their abstract ends with the bold claim that “knowledge of Latin is probably not an optimal preparation for modern language learning,” but their actual study only involved German speakers learning conversational Spanish. It seems problematic to make a blanket statement about Latin when they didn’t study English speakers attempting reading fluency in Spanish, or Japanese speakers mastering academic Italian. And since the article came out, other research has challenged their end-of-abstract declaration.

2. They actually found that Latin offered benefits to German students in linguistic activities.

They observed that German students of Latin did better on German language tests than their peers lacking Latin studies. They also noted that “Latin was associated with transfer effects on grammar-related activities.” Their titular search has apparently ended: the benefits of Latin from their own data seem to come at least in native language mastery.

3. They do not evaluate the role of Latin in sustained studies of a secondary or tertiary language.

The study at the heart of the paper involved fifty female German university students taking a fifteen-week “crash course” of sorts in conversational Spanish. All students had no Spanish experience, seven years of English (at least), and five years (at least) of either Latin or French. In other words, the students were not trying to master a sustained study of Spanish or reach for academic or literary Spanish.

4. The data was possibly tainted by the vexing presence of the participants’ instructor.

The authors reported that “76 participants of the course were informed at the end of the 13th session about a scientific study on foreign language learning being conducted…[and] also asked to complete a questionnaire” [italics mine]. In other words, their professor was seemingly present (this was not administered outside of class, or if it was, they failed to note that). This matters because the questionnaire, given without advance notice and taken seemingly in the presence or proximity of their professor, asked them to identify personal information such as age, gender, native language, and their majors. This leads into the next issue:

5. Participants were asked to self-report potentially sensitive information in proximity to their instructor.

They were also, crucially, asked to rate their interest in Spanish as a subject as well as the amount of time they had been devoting to the teacher’s material outside of class. These settings do not seem ideal for such divulging material that a student would understandably fear might sway a teacher’s opinion so near to the end of the course when students were seeking certification for job opportunities and the instructor would assess final work by hand in two weeks.

6. Participants were asked to self-evaluate their past performance in language courses without grade reporting and via a very narrow set of choices.

Arguing that students would not remember their past grades in Latin or French, the authors asked the students to subjectively rate their performance on a three-point scale (the options being “relatively good,” “average,” and “relatively bad”). This is highly subjective, and studies suggest that high achievers underrate themselves, while low achievers overrate their abilities. Without grades or testimonials from their past instructors, we really don’t know what sorts of Latin and French students participated.

7. The study’s survey of the literature somehow managed to avoid key findings in favor of Latin studies.

Their survey of the literature preceding the core of their study overlooked a number of studies from 1970s to 2003 which covered detected benefits of Latin. In other words, whereas other studies undertaken in a decade or so before the work of Haag and Stern had found numerous studies touting demonstrated benefits to studying Latin, Haag and Stern magically seemed to find mostly derogatory literature (with a key source being a century old).

8. Their presentation of Latin in comparison to Romance languages seemed intentionally poorly-constructed.

Their illustration of the connections between Latin, French, and Spanish seemed either willfully obscurantist or hopelessly unequipped to address Latin and historical linguistics. One might forgive them for ignoring the evolution of French and Spanish prepositions such as par/por and de/de from the Latin per and de, or for identifying caelum as a genitive instead of caeli. However, they made an egregious misstep by claiming that “Littera Auli became une lettre de Jacques in French and una carta de Juan in Spanish.” In English, that sentence would be “A letter of Aulus became a letter of Jack in French and a letter of Juan in Spanish.” Now, imagine if they had said “Bibliotheca antiqua ab Iulio visa became une antique bibliothèque vue par Julius in French and una antigua biblioteca vista por Julio?” Students of Latin know Latin preferred the Greek-derived epistula for postal letters (littera generally meant alphabetic letters), which survives in French as épître and in Spanish as epístola (learnèd and Biblical usages, but existent). Likewise, the Spanish carta came from the Latin charta, which could mean “paper” or “writing.” Haag and Stern seemed eager to downplay Latin’s connections to Romance tongues, and the decision to set Aulus alongside Jacques and Juan instead of connected names was at best an odd choice, at worse an intentional attempt to downplay interrelationships.

9. They didn’t reflect on the limitations of their particular study in terms of diverse student experiences.

The title of the article was “In Search of the Benefits of Learning Latin,” not “Latin’s Utility for Mastering Tertiary Conversational Discourse in Time Sensitive University Contexts,” but the latter title would have been more appropriate. How applicable are the experiences of fifty German university students with competency in three languages to those of American secondary students, particularly those from marginalized communities? They concluded by stating “specific knowledge elements and cognitive activities rather than formal rigor are transferred from one situation to another. This appears to be true for Latin.” Years later, we can note that Latin students not only post the highest verbal scores on the SAT, but often are second-highest in math scores. If we assume this is because of the likely privileged backgrounds of Latin students, we must remind ourselves that Haag and Stern found that Latin studies boosted native language proficiency. Are Latin students just math whizzes, or could Latin have contributed to the high average performance of Latin students on all areas of the SAT? The question needs better attention than works with laudatory citations of Haag and Stern can provide.

10. The final test, where the performance differences were slight, measured the ability to translate into Spanish from German, not to read Spanish or even converse in Spanish. 

This may actually give some ammunition to Haag and Stern, as Latin courses traditionally focus on reading more so than composition or conversation. However, one wonders if the students had been asked to digest a conversational or complex Spanish text if the results would have been different. It’s worth looking at the table at the bottom of their fourth page. In a couple of metrics, the French students made more errors. No discussion was given whether or not the Latin students would have received substantially worse final marks, or if what looks impressive in formal data wouldn’t have affected their ability to gain conversational Spanish certification.


It is true that proponents of Latin in the United States of America claim that studying Latin is a good preparation for studying the modern Romance languages. It is fair to note that Haag and Stern challenge this narrative, but they certainly don’t demolish the proposition alluded to by their aggressive title. If you’re against Latin studies, Haag and Stern provide you a flare gun, not heavy artillery. If you’re for Latin studies, Haag and Stern provide more evidence that Latin would benefit students, and little concrete evidence that the benefits of Latin are elusive.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s