Ancient to medieval period
Although the need to learn foreign languages is almost as old as human history itself, the origins of modern language education are in the study and teaching of Latin in the 17th century. Latin had for many centuries been the dominant language of education, commerce, religion, and government in much of the Western world, but it was displaced by French, Italian, and English by the end of the 16th century. John Amos Comenius was one of many people who tried to reverse this trend. He composed a complete course for learning Latin, covering the entire school curriculum, culminating in his Opera Didactica Omnia, 1657.
In this work, Comenius also outlined his theory of language acquisition. He is one of the first theorists to write systematically about how languages are learned and about pedagogical methodology for language acquisition. He held that language acquisition must be allied with sensation and experience. Teaching must be oral. The schoolroom should have models of things, and failing that, pictures of them. As a result, he also published the world's first illustrated children's book, Orbis Sensualim Pictus. The study of Latin diminished from the study of a living language to be used in the real world to a subject in the school curriculum. Such decline brought about a new justification for its study. It was then claimed that its study developed intellectual abilities, and the study of Latin grammar became an end in and of itself.
"Grammar schools" from the 16th to 18th centuries focused on teaching the grammatical aspects of Classical Latin. Advanced students continued grammar study with the addition of rhetoric.
18th century
The study of modern languages did not become part of the curriculum of European schools until the 18th century. Based on the purely academic study of Latin, students of modern languages did much of the same exercises, studying grammatical rules and translating abstract sentences. Oral work was minimal, and students were instead required to memorize grammatical rules and apply these to decode written texts in the target language. This tradition-inspired method became known as the 'grammar-translation method'.
19th–20th century
Innovation in foreign language teaching began in the 19th century and became very rapid in the 20th century. It led to a number of different and sometimes conflicting methods, each trying to be a major improvement over the previous or contemporary methods. The earliest applied linguists included Jean Manesca, Heinrich Gottfried Ollendorff (1803–1865), Henry Sweet (1845–1912), Otto Jespersen (1860–1943), and Harold Palmer (1877–1949). They worked on setting language teaching principles and approaches based on linguistic and psychological theories, but they left many of the specific practical details for others to devise.
Those looking at the history of foreign-language education in the 20th century and the methods of teaching (such as those related below) might be tempted to think that it is a history of failure. Very few students in U.S. universities who have a foreign language as a major manage to reach something called "minimum professional proficiency". Even the "reading knowledge" required for a PhD degree is comparable only to what second-year language students read and only very few researchers who are native English speakers can read and assess information written in languages other than English. Even a number of famous linguists are monolingual.
However, anecdotal evidence for successful second or foreign language learning is easy to find, leading to a discrepancy between these cases and the failure of most language programs, which helps make the research of second language acquisition emotionally charged. Older methods and approaches such as the grammar translation method or the direct method are dismissed and even ridiculed as newer methods and approaches are invented and promoted as the only and complete solution to the problem of the high failure rates of foreign language students.
Most books on language teaching list the various methods that have been used in the past, often ending with the author's new method. These new methods are usually presented as coming only from the author's mind, as the authors generally give no credence to what was done before and do not explain how it relates to the new method. For example, descriptive linguists seem to claim unhesitatingly that there were no scientifically-based language teaching methods before their work (which led to the audio-lingual method developed for the U.S. Army in World War II). However, there is significant evidence to the contrary. It is also often inferred or even stated that older methods were completely ineffective or have died out completely when even the oldest methods are still used (e.g. the Berlitz version of the direct method). One reason for this situation is that proponents of new methods have been so sure that their ideas are so new and so correct that they could not conceive that the older ones have enough validity to cause controversy. This was in turn caused by emphasis on new scientific advances, which has tended to blind researchers to precedents in older work.
There have been two major branches in the field of language learning; the empirical and theoretical, and these have almost completely separate histories, with each gaining ground over the other at one point in time or another. Examples of researchers on the empiricist side are Jesperson, Palmer, and Leonard Bloomfield, who promote mimicry and memorization with pattern drills. These methods follow from the basic empiricist position that language acquisition basically results from habits formed by conditioning and drilling. In its most extreme form, language learning is seen as basically the same as any other learning in any other species, human language being essentially the same as communication behaviors seen in other species.
On the theoretical side are, for example, Francois Gouin, M.D. Berlitz, and Elime de Sauzé, whose rationalist theories of language acquisition dovetail with linguistic work done by Noam Chomsky and others. These have led to a wider variety of teaching methods ranging from the grammar-translation method to Gouin's "series method" to the direct methods of Berlitz and de Sauzé. With these methods, students generate original and meaningful sentences to gain a functional knowledge of the rules of grammar. This follows from the rationalist position that man is born to think and that language use is a uniquely human trait impossible in other species. Given that human languages share many common traits, the idea is that humans share a universal grammar which is built into our brain structure. This allows us to create sentences that we have never heard before but that can still be immediately understood by anyone who understands the specific language being spoken. The rivalry of the two camps is intense, with little communication or cooperation between them.