Approaches to the use of online language tools and AI to aid language learning

Adèle Venter, Head of German at WHS, considers how, in a time when Google Translate has insidiously pervaded every homework task, students could be trained to use online language tools and AI to aid their language learning rather than lead them astray.

 

A few years ago – some of my students may still remember it – my Year 10 German class experienced a dark moment. Upon handing back their homework essays, I asked them to write me a note about the extent to which they had used Google Translate to complete their homework.

The atmosphere was grim as they sat writing their confessions.

It reminded me a bit of the confessing sheep in Animal Farm and I almost felt sorry for them. But no – this had to end. I explained to them how I was in fact not assessing their progress and understanding but rather how well (or not – as was still the case at the time) Google’s artificial intelligence manages to translate language completely out of context. I illustrated to them how they were sometimes unable to even translate the German in their essays, and how therefore, they had learnt nothing in the process, making my conscientious attempts to provide feedback on their writing a waste of time.

The Google Translate dilemma

Of course, this has been a much-discussed topic and the bane of foreign language teachers’ lives for some time now, as illustrated by this Twitter joke that did the rounds:

I still stand by everything I had said on that day. And I would like to think that it may have changed their outlook somewhat. But I have since changed my approach to it. Because, as the saying goes, if you can’t beat them, join them.

Ultimately, it is also true that the Internet has become enormously useful in helping people with language acquisition. In the first instance, various language-learning applications have seen the light of day and people casually engage with these on various levels. If it means more people are able to buy croissants in France, or have a basic conversation with their grandchildren who live in Italy, it must be a good thing, right?

Unfortunately, the one thing that has remained true for the acquisition of a foreign language is that there is no quick and easy way to do so. I am of the firm belief that to really learn a language, it takes a lot of time, dedication and perseverance, and that your best chance of becoming proficient is to combine the formal learning of its grammar and vocabulary with immersion and exposure in authentic contexts.

Can AI tools play a useful role?

And so my question is mainly: what are the implications of the use of online tools for the dedicated language learner?

As a linguist, I do not deny that I use these myself all the time. But instead of just modelling my use of online dictionaries, conjugators and such, I have decided to engage my students more fully in the conversation so that they can be conscious of the advantages and pitfalls to various tools. I have told my students that I do not consider Google Translate to be one of the seven deadly sins anymore. After all, online translators have made enormous strides in past years, and a student workshop with Mrs Rachel Evans, our Director of Digital Learning, has revealed that more often than not, they tend to translate phrases and sentences, even idioms correctly.

Instead, I spread the message that whatever students do, they must ensure that they remain in charge of the things they write down. If they do not understand what they are writing, or why sentences are formulated in a certain way, they cannot hope to learn from it. I have consequently set up the following rules as guidance:

  1. Always turn to the dictionary first. There are excellent online dictionaries, and it is worth knowing which ones can be trusted to be correct and informative. It is important that they should understand that verified dictionaries offer synonyms, context and more information about the word, which translators do not. Dictionaries are a great source for developing intuition around words in varying contexts. The more advanced student could also draw on etymology. In the making of a linguist, these are skills well worth developing.
  2. Use online technology to enhance knowledge, not replace it. If pupils use the structures they have mastered as a starting point, they could explore replacing elements of the sentence (such as verbs by researching via a dictionary or conjugator).
  3. Keep the channels of communication open. Let your teacher know how you came by a certain word or phrase. I ask my students to highlight phrases they have constructed using a translator and indicate how they researched it. What were they trying to say? Going back to my second rule of course, are there ways of bringing across their meaning, using the structures they can already manage?

At a more advanced level, language learning becomes increasingly adventurous and as students gain independence, they are able to use language tools to develop the sophistication and concision of their expression. It is mainly younger students who experience frustration around their limited ability to express themselves. The following scenario serves as the perfect example of such a problem. A multilingual girl in Year 9 who is used to expressing herself effortlessly in various languages, produces the following sentence:

„Ich liebe Little Women weil es mich zum Weinen brachte.“

I love Little Women because it brought me to tears.

“Brought” as the imperfect form of the mixed verb “to bring” was rather more than I had counted on at her level and true enough, she did not understand the verb she had used, having typed in “it made me cry”. In fact, there is a myriad of grammatical complexities in this sentence that she had not yet mastered; she could not hope to construct such a sentence with her level of skill. Instead, a well-chosen adjective in an opinion phrase would have been within her reach and might have expanded her repertoire.

Learning to be independent and in control

I hope that having an open discussion will help students to become conscious of problems such as the example shown here and encourage them to use verified sources, finding those tools that are worthwhile learning aids. If they approach it with the right mind, these tools could help them to become truly skilled linguists who are able to reflect on elements of language in a sophisticated way. If language teachers can succeed in creating such healthy learning habits, they are likely to make a meaningful contribution towards developing students’ independence and ability to be life-long learners in the age of technology.

Who is in control? The human being.

Speaking in tongues: why reconstruct a language we don’t even know existed? – 09/11/18

Anna (Year 13) looks back to our earliest beginnings as a civilisation in the Indo-European world, discovering that there is only one route to the reconstruction of Indo-European culture that offers any hope of reliability and that is language.

Swedish, Ukrainian, Punjabi, and Italian. To many of us, these languages are as different and distinct as they come. But it has been discovered that, in the same way that dogs, sheep and pandas have a common ancestor, languages can also be traced back to a common tongue. Thus, Dutch is not merely a bizarrely misspelled version of English and there is more to it than our languages simply being pervaded by the process of Latin words being imported into native dialects in the Middle Ages.

In the twelfth century, an Icelandic scholar concluded that Englishmen and Icelanders ‘are of one tongue, even though one of the two [tongues] has been changed greatly, or both somewhat.’ He went on to say that the two languages had ‘previously parted or branched off from one and the same tongue’. Thus, he noticed the common genetic inheritance of our languages, and coined the model of a tree of related languages which later came to dominate how we look at the evolution of the Indo-European languages. We call this ancestral language Proto-Indo-European, a language spoken by the ancestors of much of Europe and Asia between approximately 4,500 and 2,500 B.C.

The Indo European Family Tree

But what actually is it? Well, let me start simply. Consider the following words: pedis, ποδος (pronounced ‘podos’), pada, foot. They all mean the same thing (foot) In Latin, Ancient Greek, Sanskrit and English respectively. You will notice, I hope, the remarkable similarity between the first three words. English, on the other hand, sticks out slightly. Yet, it has exactly the same root as the other three. If I were to go back to one of the earliest forms of Germanic English, Gothic, you may perhaps notice a closer similarity: fotus. Over time, a pattern emerges: it is evident that the letter p correlates to an f and a letter d to a t. This is just one example of many: it is these sound laws that led Jacob Grimm to develop his law.

Grimm’s law is a set of statements named after Jacob Grimm which points out the prominent correlations between the Germanic and other Indo-European languages. Certainly, single words may be borrowed from a language (like the use of the words cliché, from the French, or magnum opus, from Latin), but it is extremely unlikely that an entire grammatical system would be. Therefore, the similarities between modern Indo-European languages can be explained as a result of a single ancestral language devolving into its various daughter languages. And although we can never know what it looked like, we can know what it sounded like. This is because, using Grimm’s Law, we can construct an entire language, not only individual words, but also sentences and even stories.

In 1868, German linguist August Schleicher used reconstructed Proto-Indo-European vocabulary to create a fable in order to hear some approximation of PIE. Called “The Sheep and the Horses”, the short parable tells the story of a shorn sheep who encounters a group of unpleasant horses. As linguists have continued to discover more about PIE, this sonic experiment continues, and the fable is periodically updated to reflect the most current understanding of how this extinct language would have sounded when it was spoken some 6,000 years ago. Since there is considerable disagreement among scholars about PIE, no single version can be considered definitive: Andrew Byrd, a University of Kentucky linguist, joked that the only way we could know for sure what it sounded like is if we had a time machine.

The earliest version read as follows:

(The audio of a later version, read by Andrew Byrd can be found at the following link: https://soundcloud.com/archaeologymag/sheep-and-horses)

Here is the fable in English translation:

Though seemingly nonsensical, it is definitely exciting, and when you take a metaphorical microscope to it, you can notice similarities in words and grammar, particularly that of Latin and Ancient Greek. What is the point, though, in reconstructing a language no longer spoken?

Firstly, the world wouldn’t be what it is today had it not been for the Indo-Europeans. If you’re reading this article, chances are that your first language is an Indo-European language, and it’s also very likely that all of the languages you speak are Indo-European languages. Given how powerfully language shapes the range of thoughts available for us to think, this fact exerts no small influence on our outlook on life and therefore, by extension, on our actions.

Secondly though, as a society, we are fascinated by our history, perhaps because examining our roots (to continue the tree metaphor) can help us understand where we may be headed. Although many archaeologists are hesitant to trust linguistic data, by gaining an insight into the language of the PIE world, we can make inferences about their culture and in turn learn more about our own. One such example of this is Hartwick College archaeologist David Anthony’s discovery of a mass of sacrificed dog and wolf bones in the Russian steppes. By consulting historical linguistics and ancient literary traditions to better understand the archaeological record, he and his team found that historical linguists and mythologists have long linked dog sacrifice to an important ancient Indo-European tradition, the roving youthful war band (known as a ‘koryos’ in reconstructed PIE). This tradition, which involved young men becoming warriors in a winter sacrificial ceremony, could help explain why Indo-European languages spread so successfully. Previous generations of scholars imagined hordes of Indo-Europeans on chariots spreading their languages across Europe and Asia by the point of the sword. But Anthony thinks Indo-European spread instead by way of widespread imitation of Indo-European customs, which included, for example, feasting to establish strong social networks. The koryos could have simply been one more feature of Indo-European life that other people admired and adopted, along with the languages themselves. We can learn about the customs of our prehistoric ancestors and so Indo-European studies is relevant because as powerfully as it has influenced our modern social structure and thought, there are also many ways in which the Indo-European worldview is strikingly different from our own. Studying it enables you to have that many more perspectives to draw from in creating your own worldview.

National Historical Museum Stockholm: A bronze Viking plate from the 6th century A.D. depicts a helmeted figure who may be the god Odin dancing with a warrior wearing a wolf mask.

The Theory of Deconstruction – 21/09/18

Ava (Year 13, Head Girl) explores the Theory of Deconstruction as suggested by Derrida and discusses the confusing nature of both ideas and words.

Deconstruction is a theory principally put forward in around the 1970s by a French philosopher named Derrida, who was a man known for his leftist political views and apparently supremely fashionable coats. His theory essentially concerns the dismantling of our excessive loyalty to any particular idea, allowing us to see the aspects of truth that might be buried in its opposite. Derrida believed that all of our thinking was riddled with an unjustified assumption of always privileging one thing over another; critically, this privileging involves a failure to see the full merits and value of the supposedly lesser part of the equation. His thesis can be applied to many age-old questions: take men and women for example; men have systematically been privileged for centuries over women (for no sensible reason) meaning that society has often undervalued or undermined the full value of women.

Now this might sound like an exceedingly overly simplistic world view, and that Derrida was suggesting a sort of anarchy of language. But Derrida was far subtler than this – he simply wanted to use deconstruction to point out that ideas are always confused and riddled with logical defects and that we must keep their messiness constantly in mind. He wanted to cure humanity of its love of crude simplicity and make us more comfortable with the permanently oscillating nature of wisdom.  This is where my new-favourite word comes in: Aporia – a Greek work meaning puzzlement. Derrida thought we should all be more comfortable with a state of Aporia and suggested that refusing to deal with the confusion at the heart of language and life was to avoid grappling with the fraught and kaleidoscopic nature of reality.

This cleanly leads on to another of Derrida’s favourite words: Differánce, a critical outlook concerned with the relationship between text and meaning. The key idea being that you can never actually define a word, but instead you merely defer to other words which in themselves do not have concrete meanings. It all sounds rather airy-fairy and existentialist at this level, but if you break it down it becomes utterly reasonable. Imagine you have no idea what a tree is. Now if I try and explain a tree to you by saying it has branches and roots, this only works if you understand these other words. Thus, I am not truly defining tree, but merely deferring to other words.

Now if those words themselves cannot be truly defined either, and you again have to defer, this uproots (excuse the pun!) the entire belief system at the heart of language. It is in essence a direct attack on Logocentrism, which Derrida understood as an over-hasty, naïve devotion to reason, logic and clear definition, underpinned by a faith in language as the natural and best way to communicate.

Now, Derrida clearly wasn’t unintelligent, and was not of the belief that all hierarchies should be removed, or that we should get rid of language as a whole, but simply that we should be more aware of the irrationality that lies between the lines of language, willingly submit to a more frequent state of “Aporia”, and spend a little more time deconstructing the language and ideas that have made up the world we live in today.

Taking the register

James Courtenay Clack considers the way we use language in the classroom.

It started with a pupil in my Year 9 class dismissing Romeo and Juliet’s first meeting as ‘flanter’. Whilst inwardly I shared her sniffiness about their saccharine shared sonnet, there seemed to be a clear challenge coming from this pupil. Would I, as a teacher, and an English teacher no less, allow her to use slang to talk about one of the most famous scenes in all of literature? This got me thinking about the way we use language in the classroom.

With morning and afternoon registration and seven lessons, most pupils can expect to answer the register nine times a day. As important as this rite is in keeping track of pupils, far more interesting is the other type of register pupils flit between. If anyone with a half-keen ear were to follow a pupil around the school for a day, they would notice that the way pupils speak and write changes as they move from home to school, from WhatsApp groups to essays. Rare is the pupil, for example, who answers the question of what they want for breakfast in the same formal and detailed manner in which they would a question about the bleak landscapes of TS Eliot’s ‘The Waste Land’.

Register, as coined by Thomas Bertram Reid and developed by Martin Joos, is the sociolinguistic term for the use of language as defined by social situation. Broadly speaking, register is about the level of formality of language and it incorporates elements as disparate as:

  • Vocabulary
  • Tone
  • Dialect
  • Slang
  • Abbreviation
  • Use of full sentences

It is worth thinking about why we use language differently in different contexts. In certain social situations, the answer is obvious. We adopt a more formal register with managers/bosses at work, for example, both as a sign of respect and out of a desire to impress. Similarly, our use of a more informal register with friends belies a level of comfort and intimacy that we do not share with our employers.

For the most part the ability to move between these registers – or at least the knowledge that you speak differently with Mrs Lunnon than you do with your BFFL – is picked up when young. As our social skills develop, our ability to move between registers becomes almost unthinking. When I applied for a job at WHS, for example, I did not have to remind myself not to greet Mrs Lunnon with an ‘alreet pal’ at the start of my interview.

That we use language differently in different social contexts is hardly a ground-breaking observation, but it does have several implications for teachers. Once pupils enter the classroom they are introduced to a new range of registers and in English, there are two main areas of interest, or areas of clash: the way we talk in class discussions and the way we write essays.

Although essays are the main form of assessment, the primary skill being assessed – interpretation of a literary text – is developed in class discussion and debate.

The question of how we speak in class is deeply political. The fact that the ability to use language in certain ways is a form of social currency in this country has moved schools such as Michaela Community School and Harris Academy Upper Norwood to ban the use of certain registers in the classroom.

The desire to equip pupils with the verbal skills required for social mobility is undoubtedly a noble one, but I would argue that such a hard-line approach to register is dangerous. Firstly, the fact that WHS’ A-Level curriculum includes writers as diverse as Chaucer and Tennessee Williams is proof enough that there isn’t one ‘correct’ form of English. Secondly and more importantly, however, is the entire point of class discussions. The reason we discuss and weigh up ideas and not just dictate, Mr Gradgrind style, from the front is that we want to encourage pupils to develop their own voices and their own thoughts.

If pupils are going to develop ideas and formulate opinions through discussion, as they will have to do at university, they will need to feel a sense of ownership over the curriculum. It is for this reason that I encourage my class to voice opinions about texts in their own way and it was for this reason that I whole-heartedly agreed with my Year 9 pupil about Romeo and his ‘holy shrine’. Similarly, if my Year 12 class want to refer to the author of Mrs Dalloway as Ginny Bae, or if my Year 8s want to describe Shylock’s insistence on his pound of flesh as ‘a bit extra’, then why should they not?

It is possible to create an enjoyable learning environment in which pupils feel confident voicing ideas and opinions in their own language whilst at the same time modelling the more rigorous language of academia. So I agreed that yes, Shylock ‘was a bit extra’, but then questioned them as to whether his desire for revenge along with Antonio’s overt anti-Semitism made Shylock a tragic hero in the eyes of a modern audience or whether he was a mere comic foil about whose suffering a contemporary audience would not have cared a jot. That he is both, and more, is testament to the genius of Shakespeare, a man who knew more than most about the power of mixing the language of the court with that of the street YEAH.

Twitter: @English_WHS