I have entitled this paper a ‘manifesto’ because I intend to articulate my thoughts in a declamatory, and possibly inflammatory, manner, without recourse to extensive documentation. I have chosen the phrase ‘classical languages’ with equal deliberation. By this phrase I mean those languages that have ceased in their ancient standards to be used by a continuous population as a first language for everyday communication. Thus, although Latin has evolved into a host of Romance languages, those languages are not Latin per se, and likewise Modern Greek is significantly diverse from the Greek that came to revolve around an Attic standard of the 4-5th centuries BC. The two halves of this phrase exhibit peculiar features to which I wish to draw attention. ‘Languages’ which are ‘Classical’
Firstly, the object of our discussion is ‘languages’, by which I mean something like ‘self-contained culturally-entwined systems of communication’. I draw particular attention to the fact that they are self-contained. By this I do not mean they exist without reference to an ‘external world’, but rather they do not rely or need an external system of communication, such as another language, for them to function. In this they differ from a ‘code’ or ‘cipher’, which in fact must be interpreted with regard to an external system.
Secondly, I use the term ‘classical’ in preference to two alternatives, ‘ancient’ and ‘dead’. ‘Ancient’ unnecessarily clouds our perception of the corpus of linguistic usage under consideration. Latin, for example, was used virtually as a lingua franca well into the 17th, 18th, and in some situations 19th century. Attic Greek remained a standard well into the Byzantine period. ‘Ancient’ in these cases fails to do justice to the breadth of temporal usage. ‘Dead’ is a pejorative term, implying that they are languages that are useless for active, regular, and everyday usage as communicative systems. In the classroom, all languages are ‘dead’, and there is nothing intrinsic to the languages under discussion that makes them any more ‘dead’ than any other synchronic snapshot of any other language. The designation ‘classical’ serves a number of useful purposes. It draws our attention to the fact that these are languages which regularly have a grammatical ‘standard’ to which a vast body of literature attempts to emulate as ‘correct’. In the case of Latin, that standard tends to be 1st Century BC Latin, and Ciceronian prose in particular. In the case of Greek, it is the Attic standard which dominates from the 5-4th centuries BC in Athens, and extends into the Hellenistic and Imperial periods, and well into the Byzantine period as the standard by which literature is measured. A similar case can be made for a number of similar languages (‘Classical’ Hebrew, ‘Classical’ Arabic, etc..). It also draws attention to the fact that the main purpose of learning these languages is usually to engage in a communicative act (primarily reading) with the ‘literature’ (often literature of ‘high’ artistic value, though not necessarily so restricted) of these cultures, especially that body of literature that has been canonised, either by those cultures or their cultural heirs. However much we may, and often rightfully, critique notions of ‘literature’ and ‘canon’ and ‘classical standards’, such notions are historical realities that must also be grappled with in these fields of study. ‘Language Acquisition and Linguistic Knowledge’
I now wish to articulate a distinction I derive primarily from Stephen Krashen, inter alia. This is the distinction between what I will term here ‘Language Acquisition’ and ‘Linguistic Knowledge’. By Linguistic Acquisition, I refer to the gradual progress in learning a language as a means of communication and thought. The best way to understand this is to think of your first, or native, language. This is, for most, the language in which they think, and the language in which they articulate their thoughts, etc., to others in communicative acts (speech, writing), and understand others so communicating (hearing, reading). To acquire a second, or subsequent, language (except in the case of infant bi- and multi-linguals) involves a process by which the learner gains, or acquires, the ability to perform those functions with a separate (though sometimes related, as languages are) system of communication independently of their prior language(s). That is, we refer to someone as ‘knowing’, for example, Japanese, if we consider or observe them capable of communicating (reading, writing, hearing, speaking) with that language.
‘Linguistic Knowledge’ properly refers to knowledge about a language. It is what is often termed ‘grammar’. It is possible, of course, to learn the grammar of one’s native language, and to study the grammar of other languages. That linguistic knowledge does not equate with language acquisition is readily, I believe, demonstrated by considering the training of linguists. A well-trained linguist is inevitably acquainted with a number of languages. They also have a solid grasp of linguistics, and usually the ability to understand the structures and meta-structures of language. Give a linguist a grammar (i.e., a description of the structure of a language), a lexicon (a set of glosses or interpretive entries for the lexemes of a language), and a sample text, and given appropriate time they can produce an accurate translation, or obtain an accurate understanding, of that text and its meaning. We would be odd, indeed, if we considered such an ability to be the same as ‘knowing a language’. We would not normally conclude that such a linguist ‘knew’ that language.
The development of the Grammar-Translation methodology
Yet, when many come to consider classical languages, this is precisely the mistake they make. I believe a historical argument may here make my point abundantly clear. With the rise of vernaculars, i.e.., the Romance and other European languages, in Europe, Latin became increasingly used in smaller and more ‘elite’ circles – nobility, academics, ecclesial practice. This led, in turn, to a decreasing facility with the language, both by its practitioners and by its students. The force of this decreasing facility created a kind of downward spiral of pressure – it made it harder to use, so it was used less. Using it less made it harder to use, and even harder to acquire. A second downward spiral attached to the first – the argument of utility. As Latin declined, vernaculars increased in use and importance, so that the need to learn Latin became less obvious and pressing. In turn, despite certain conservatisms, the standards for Latin acquisition became less taxing – the removal of spontaneous oral practice, then prose composition, and so on, until translation from Latin texts into native languages was the main and only practice of so much Latin study. This decline was also impacted by what I term the rise of ‘hyper-grammar’ in Germany. German linguistic studies came to focus on a rigorous analysis of grammar and syntax in texts, with elaborate schemes of classification. These studies, academically, were invaluable, and from a certain point of view I am very grateful for this academic ‘movement’. Its impact pedagogically, I believe, is almost disastrous. The need to produce students capable of such detailed and sophisticated analyses lead to the rise, and triumph, of grammar-translation as a pedagogic method for the teaching of languages. Now, to teach ‘Latin’ was to equip students with a detailed knowledge of the grammar of Latin, and instil in them a vast-enough knowledge of Latin vocabulary. This was, essentially, to train linguists (See above) specialising in Latin. Grammar-Translation has largely fallen away as a method in most modern manguages, but classical languages have moved relatively little on pedagogic methodology over the past 150 years. The Failure of Grammar-Translation
Now, I wish to make a claim which I believe will be offensive to many, and damning to grammar-translation as a methodology. Grammar-Translation only produces persons who have ‘acquired’ a classical language as an accidence, not a result, of its methodology. Furthermore, Grammar-Translation is entirely, theoretically if not yet practically, replaceable by computerised processes.
I will deal with the latter point first. It is true that at the time of writing computer translations are quite dismal. However, I believe the challenge is technical, not innate. Grammar, or syntax, is largely a system of rules. Computers are designed to handle rule-based systems. Lexicons, likewise, are largely compilations of appropriate glosses, or interpretations, from one language to another. The resistance to computer translation is primarily predicated on the ‘nuance’ and ‘feel’ that a human translator brings to the task – they understand what nuance or relevant gloss a word deserves in context, and how the structure operates in context. I suspect that human resistance to the idea of computer-translation replacing human-translation is part of an unwillingness to recognise that this is not necessarily about ‘fuzziness’. What would tip the balance would be the ability of computers to contextualise structures and morphemes properly. This is in fact easier to imagine with classical languages, which tend to be a closed corpus of texts. TLG and other such resources have digitised such corpora. A properly designed program with enough computing ‘power’, could manage the task of reading a text’s words and structures and comparing them to relevant parallel structures and semantic contexts, to produce what we would consider fair translation.
If this is true, it suggests that the work of highly trained human students of classical languages is really the internalisation of something akin to a mechanical process. That is, we expect students of classical languages to internalise the grammar and internalise the vocabulary (i.e., we really expect them to learn rules and meanings by rote), and then produce translations. It is really only the rote internalisation that sets them apart from a trained linguist, not a trained classicist, who could come and do the same process, albeit far slower.
Now, I will not yet return to my claim that the products of Grammar-Translation methodologies only acquire classical languages as an accident, not a result, of their methodology. To come to that point I need to articulate a case for Language Acquisition as quite a different phenomenon. A picture of Language Acquisition
The nature and processes of (Second) Language Acquisition remains a hotly debated field. I agree, in broad brush strokes, with the theories put forward by Stephen Krashen et alia. I am going to try and articulate here a more generic description of language acquisition.
To acquire a language, that is to gain a competency in using it as a means of communication without recourse to another language, is a process. It’s a process that is largely defined and constrained by comprehensibility. Let’s take an hypothetical example.
Learner A has only a basic ability to comprehend Latin. They have taken a day or so of TPR1 teaching. Yesterday they learnt that ‘da mihi aquam’ is roughly equivalent to ‘give me water’. They learnt this by physical demonstrations and gestures, and came to understand that ‘da’ is an utterance that orders the addressee to give an object. They learnt that ‘mihi’ is an utterance, which in this context, indicates that the object should be given to the speaker. In fact, they learnt this in contrast to another utterance, ‘ei’ which they understood to direct them to give the object to a third party. They also learnt that ‘aquam’ refers to water, when water is the object of a verb. Today, Teacher B brings in a number of new objects, including some wine. Teacher B uses Learner C, and firstly gives them the cup of water and says ‘da mihi aquam’ and waits expectantly for Learner C to hand over the water. Then, substituting the wine for the water, says ‘da mihi vinum’ and waits equally expectantly.
This is a most basic description of a TPR-like scenario. In it, Learner A demonstrates a comprehension of ‘da mihi aquam’ and its constituent parts, in a certain set of syntactical contexts. They also have the means to understand how ‘vinum’ will fit in, and because of this learns that ‘vinum’ refers to wine.
This illustrates the comprehensible input hypothesis in its broadest terms, that learners acquire language when exposed to language of a type where a learner knows (n) language and is exposed to (n + 1) language, in the most abstract terms. If they are able to comprehend the rest of the language structures and meaning, and have enough context, whether physical, ‘textual’, or even translation, they will be able to comprehend, within the language itself, the meaning of either the structure or lexeme that is the ‘+1’.
I believe this is a basic, general, yet accurate description of how language acquisition really works, from the earliest stages to the most advanced. In the earliest stages, pedagogically speaking, it works best if they input is tightly controlled, in the latest stages no such control is so necessary, since the amount that is comprehensible of the context both numerically and with regards to ratio, is large enough generally to ensure comprehensibility. Natural and Direct Methodologies
At this stage, I will divert and make some claims about the Direct Method, the Natural Method, and my own beliefs about language acquisition.
The Natural method and the Direct Method are often conflated, and I wish to begin this subsection by making a clear terminological distinction. By the Direct Method I refer to a system of instruction designed to introduce a student from the beginning to work in, and as far as possible only in, the target language. The Direct method, in this regard, is similar to ‘Immersion’, though Immersion often means ‘to immerse in a context of the target language without regulating the nature of the input at all’ – as, for example, a foreign exchange student with no prior instruction in the language of the culture/country to which they are exchanging. The Direct Method, in contrast, is built around a controlled and measured introduction of new material, with the aim that new material can be understood by observing its contextual introduction, grasping the previous material introduced, and observing the differences (for language systems are built on differentiation).
The Natural method is often confused with the Direct Method, and both are criticised along the following lines. The criticism is levelled that they refer uncritically to the way a child learns a language, without recourse to another language, and that this referral is invalid, since adult, or even adolescent learners, ‘as well we know’ do not learn as children do. I want to affirm this criticism where it applies. Second Language Acquisition theory has at least, I believe, a consensus on this – young children learn languages differently to us. Properly speaking, I apply the term ‘Natural Method’ to methods which aim to ape childlike acquisition of language. The Direct Method(s), distinct from this, understand and seek to properly apply, the distinctive differences in language acquisition of adolescent and adult learners.
To this end, a well designed Direct Method course: (a) works, so far as possible, in the target language, (b) is structured to take account of the faculties and abilities of adolescent and adult learners, (c) introduces both lexemes and syntactical structures in a way that ensure their comprehensibility, (d) moves steadily from (n) to (n + 1), and repeatedly.
The result of this, or any successful method of language acquisition, is a facility to communicate in the target language. For scholars of classical languages, this most often revolves around an ability to read texts in the classical language, and to be properly called language acquisition, to read such texts in the classical language with comprehension without recourse to a second language. The accidental products of extensive Grammar-Translation
Now I may return to my contention regarding the product of Grammar-Translation. The fact is that extensive reading is one form of input. Grammar-Translators make their reading comprehensible by translating the text into another language. This, while I believe substandard, is not a failure to make comprehensible. That is to say, to translate a text or utterance or whatever, while from my position pedagogically inferior and even debilitating, is still one way to render that text comprehensible. Now, given enough reading, even translating, Grammar-Translators over time will (however painfully and slowly) expose themselves to a wealth of comprehensible input. Thus, and this pertains I believe even more directly to products of American-style programs, graduates of doctoral programs in particular, required to read extensive bodies of literature in classical language(s), often2 become adept readers (if never listeners, speakers, or writers) of those classical languages. That said, there are indeed no small number of graduates at all levels who find themselves not readers of classical texts in classical languages, but highly trained linguists translating from an internalised grammar and lexicon, and thinking and working entirely in a second language. Contemporary courses
With regards to contemporary available courses or materials dealing in language acquisition, while I appreciate the great advance that ‘reading’ methods have made (see, for example, the Cambridge and Oxford Latin Courses, the JACT Greek Course, even Athenaze for Greek tends in this direction), I believe these fall well short of the kind of materials that a Direct Methodology might produce, and ultimately the kinds of standards I will articulate further below. I personally consider a number of resources to remain outstanding in their field at this time.
These include Hans Ørberg’s ‘Lingua Latina per se Illustrata’3, an inductive Latin course in two parts, which strictly and effectively builds comprehension without recourse to an external language, working solely with Latin and pictures. Nor is LL grammar-deficient, teaching Latin grammar in Latin terminology. With regards to biblical languages, Dr. Randall Buth’s materials (and, I would expect, ulpans) for Koine Greek and Biblical Hebrew, beginning with audio materials linked to pictures, only later moving on to writing, and designed to work primarily in the target language.4 Dr. Buth pays his acknowledgements to Harry Winitz’s ‘The Learnables’ materials, which I am only acquainted with via samples, but seem to follow a similar line.5 Evan Millner’s ‘Latinum’ project6, while expansive and nobly aimed, I would not classify as a direct method per se, though its aim is properly Latin known as Latin and as a language. I will return below to a consideration of the aims, methods, and outcomes of what I would describe as an ideal program of language instruction. Outcomes of a tertiary level language program
At this point, I address my thoughts to the Academy and the Church, more specifically ‘the Seminary’, and ask the question, ‘what sort of result would we want from a 3-4 year language program in a classical language?’ My answer to such a question is as follows:
A person who can read with considerable ease texts from the corpus of the relevant classical language, without recourse to a second language
A person who can apply both linguistic and literary7 levels of analysis to such texts
A person who has reached a competency level in both those skills, which will enable them to do so post-graduation, maintained by no further activity than the actual practice of doing so.
That is, I envisage that graduates would leave a program both having acquired a language to a certain level of proficiency, and having learnt about a language, that is a linguistic knowledge of the features of the language, and be able to combine both skills in practice. In the case of seminary graduates, I have in mind the regular reading of Scriptural texts, exegetical practice, hermeneutics, and the preparation of sermons, etc.. For a university graduate in a classical language, they should be able to read and study texts in the language they have studied, at roughly a sustainable level if they continued to read the kinds of texts they were reading in their final year (and obviously with more advanced skills and competency if they were to pursue further study).
Such a program must, by nature, by hybrid to some extent. It is not teaching, as some Modern Foreign Language programs have the capacity, students to be able to order cafe lattes in Rome in Latin, for whatever purpose that might serve. Neither, should it, leave students unable to do so. On the other hand, it must equip students to critically and linguistically analyse their texts, lest like so many high-school graduates who speak ‘English’ or whatever passes for it, they lament not knowing what a ‘verb’ or ‘noun’ is. Helpful Benchmarks
At this point, I suggest a couple of benchmarks which may be helpful in imagining what this might look like. Firstly, in recent years Zondervan has produced both ‘A Reader’s Greek New Testament’ and ‘A Reader’s Hebrew Bible’, and Hendrickson has produces ‘The UBS Greek New Testament: A Reader’s Edition’.8 Both include contextualised glosses for low frequency occurring vocabulary. With studies such as Diederich’s9, Major’s10, and the like, demonstrating that the frequency of vocabulary items in the corpora of classical languages has certain critical-mass points (particularly the 80% mark) and diminishing returns, a good case can be made that facilitating reading by such texts is pedagogically useful, sound, and in terms of ongoing usage of language acquisition, almost indispensable.
Secondly, the guidelines of the Common European Framework of Reference for Languages (CEFR), while abstract, provide another useful benchmark type approach. I recall something to the effect that there is marked gap between Modern Foreign Language programs among students who achieve a B2 level competency, and those that manage a C1 level as an outcome of their course. The acquisition of vocabulary
The nature of having closed corpora of texts in classical language studies leads me to suggest that four bodies of vocabulary need to be built-in to any language program: (1) high-frequency vocabulary from the classical corpus, (2) high-frequency vocabulary from contemporary first-language usage, (3) classroom-specific meta-language, and (4) grammatical terminology.
The importance of (1) should be obvious. If our ultimate aim is to train students to read texts in classical languages, they need a competent comprehension of the vocabulary of those texts. (2) is far less obvious, and I will make my case for it here. Language is acquired and sustained best by usage, and while our emphasis may be reading, I believe it is a foolish neglect to focus solely on reading. The nature of language is that it exists in (often discreet) two systems, an oral/aural, and a literary form. By teaching classical languages as ‘active’ language, we engage more of the student, and more of their senses, and more of their learning faculties, more of the time, in more ways. If we, furthermore, do so with a vocabulary that relates to everyday life in the here and now, i.e.. contemporary vocabulary for contemporary life, we create the possibility, if not the reality, that students, and even more so if we can foster a sense of linguistic community among a cohort of students, will use the language outside the classroom, thus revivifying the language, if only in part. The ultimate aim of this is not to have them speak the language for its own sake, but its pedagogic value in their acquisition of the language, I believe, will far outstrip any ‘lost time’ spent in ‘extraneous trivialities’.
Class (3) of vocabulary consists of phrases that keep a classroom functioning in the target language even as they communicate about the class. Phrases like ‘quomodo dicitur ‘apple’ latine?’ and ‘quod significat ‘malum’?’, and ‘aperite libros tuos’ all fulfil that function.
Finally, class (4) of vocabulary offers a unique possibility. So often the choice may seem to be ‘grammar’ or ‘language’, but for many classical languages the users of those languages have long ago done grammatical work on their own language in their own language. There is no obstacle, then, to teaching the grammar of those languages, at least primarily, in the terminology of the target language itself. This is, for example, the approach of Hans Ørberg in his Latin materials. This serves a number of functions - it keeps students working in the target language, even as they talk about the target language. It raises, further, the prospect that academic discussion and papers on classical languages for scholars of classical languages might actually be conducted in classical languages. Further, it provides the possibility that our understanding of classical languages might be informed by how the users of those languages themselves understood their grammar. We may, and quite rightly, come to disagree with their own understandings and systematisations, but they deserve a voice in our studies. A Way forward
Lastly, then, let me paint a vision for the way forward. What might a course of study with such aims, methodology, and hopes look like?
I believe such a course would have the following characteristics: It would work, ab initio, in the target language
It would involve exposure to both written and aural input
It would begin with pictures, actions, etc., to facilitate a commencement in the target language
It would build systematically from the known to the unknown, using context and any other aids to render new input comprehensible at every level
It would teach to a classically recognised standard for the language, and explore developments, dialects, and unusual features later in the course
It would build towards and incorporate extensive reading and listening
It would employ measures to check comprehension by means of the target language (e.g., questions in the language, etc..) rather than by translation
It would incorporate over time four classes of core vocabulary: Frequently occurring terms from the classical corpus
Frequently occurring terms from daily life, including appropriate neologisms
Usable classroom discourse
Grammatical terms from the target language
It would aim over 3-4 years to produce students who: Could function at a CEFR level of B2 or above in all four language skills: reading, writing, listening, speaking
Could speak critically about the language in the language
Had read a broad spectrum of the classical corpus
Could continue to read, analyse, and study texts from the classical corpus
Could maintain (d) as a practice, with appropriate tools, primarily through the practice of the same.
Could use the language in daily discourse
Could perform (a-f) without necessary recourse to another language
I have written this paper in part to give written shape to thoughts that continue to occupy my mind. There are some who will read this and move to dismiss it, to whom I ask, think again. I have written without argumentation, but not without a clue, and come from a background in both Classics and Theology, as well as Literary Theory, and have read academically in the fields Linguistics and particularly Second Language Acquisition. I have spent considerable time studying, learning, and reflecting on a number of languages, classical and modern, under a number of methodologies.
There are others who will read this, and perhaps be prompted to sharp disagreement rather than dismissal. I implore us all to a lively and rigorous debate. Let’s not leave these issues alone.
There are some who will read this and find some of their own thoughts, and practices even, staring back at them. I would call for greater collaboration, for teachers and professors to consider research and training in language teaching.
Lastly, there are, no doubt, students who will find themselves, like me, greatly respecting their learned instructors, but quietly longing for something more. I have spent myself countless hours studying, seeking out better methods, supplementary materials, anything to fill these gaps, and have not entirely succeeded. I don’t have all, or even many, of the answers, but I’m hopeful that the future of classical language instruction, and thus studies in classical languages overall, can be changed for the better.
Sydney, Australia, 2008.
Private correspondence can be addressed to firstname.lastname@example.org
1 TPR – Total Physical Response. A teaching methodology devised and popularised by James Asher. A development of the same is TPRS (Teaching Proficiency through Reading and Storytelling).
2 I almost say ‘inevitably’, since the sheer quantity of reading has to have some effect, I would hypothesise. Nonetheless, so long as it is translation work, even if only mental-translation work, I stand by ‘often’.
7 Such as, for example, ‘Discourse Analysis’
8 The UBS version is a welcome addition, the Zondervan GNT being based on the ‘text behind the NIV’, less than satisfactory I would argue for most seminary graduates. Now, if only they would improve their font selections.
9 Diederich, P.B. ‘The Frequency of Latin Words and their Endings’ The University of Chicago Press: Chicago Illinois, 1939. http://users.erols.com/whitaker/freq.html
10 Major, W.E. ‘It’s not the Size, It’s the Frequency: The Value of Using a Core Vocabulary in Beginning and Intermediate Greek’ CPL Online Vol 4, No 1, Winter 2008. http://www.camws.org/cpl/cplonline/cplonline.html