Lingua.ly offers learners an opportunity to look up and save words they come across when reading online articles in another language and then recommends relevant texts for the learners to read in order to extend their vocabulary and improve their reading skills.
Users can click on any word in an article they are reading online and, after showing a definition and playing an audio version of the word, the system stores the word in a user’s personal word collection or ‘inbox’. These words can then be practiced and repeated with activities that ask users to match the words with their translations. Practice sessions are available in 5, 15 and 30 minute chunks depending on how much time the user has available and the system recommends that users try and keep their inboxes clear (i.e. regular practice).
Lingua.ly stores data about the words that the user knows and how well and then uses that data to suggest texts that the user should read in order to reinforce and extend their lexical knowledge. All texts are also tagged and categorised so that learners can filter by topic or area of interest. The more time a user spends on the system, the better the recommendation that the system makes.
The system allows users to store words in English, Spanish, French, Hebrew and Arabic and works either as a browser plugin, a web app or android app.
One of the biggest plus points of the Lingua.ly platform is the amount of comprehensible input that the learners are exposed to. The algorithm builds up an understanding of the words that a user knows and aims to suggest articles to learners where they have around a 90% text coverage, allowing for acquisition of new words from context and with the help of the app. Although, it has to be said that some of the article choices are quite strange, maybe this would improve as the the system learned more about my profile.
Another strong point on the educational side is that of autonomy. The app allows user to take control of their learning by selecting the words that they would like to learn and read the texts that they would like to read. At the same time there is enough scaffolding and support to mean that the amount of control doesn’t put a significant stress or excess of cognitive load onto the learner. There is always the option to go back to a recommended text or to practice the saved words in familiar exercises.
This brings us to another excellent aspect of Lingua.ly, the opportunities for practice. The algorithms try to suggest articles which contain some of the words you have saved in the last 72 hours, in order to reinforce them. Plus spaced repetition algorithms take into account when you’ve seen the word, how well you know it and its connectedness to other words you know in order to decide how often you need to be tested on the word to maximise the chance of committing it to long-term memory.
The price, at completely free to all users (currently), is also an impressive selling point! This does mean that the translations that you see of a word are occasionally suspect (more on that later), and therefore you may be tested on these strange translations of saved words at a later date.
With so much reading happening online and on mobile these days, Lingua.ly acts as an important support for a real, everyday task that language learners face. Indeed the founders themselves learned langauges by manually following the Lingua.ly process of looking up words in articles and then using spaced repetition software to help them memorise the unknown words. Lingua.ly was created in order to make this process more efficient, and through article recommendation, more effective too. By helping with a very definite real word need, it’s likely that users will find the product helpful and use it enough to see some real benefit.
And finally, from looking at the team involved in the product, it is clear that they have an impressive amount of background knowledge and experience, including expertise in SLA, natural language processing, computational biology and cognitive psychology. And from looking at the Lingua.ly blog it’s clear that there is a lot of knowledge of teaching and learning and a real desire to help learners aquire language effectively and efficiently.
The main weakness, given the intended outcomes of the product, is the lack of chunking of the lexis. All words are treated individually, with no work on phrases, formulaic language or collocation. So when you click on a word to save it, the translation engine offers many different translations but automatically saves the first one, regardless of part of speech or the specific sense of the word (e.g. clicking on Spanish ‘general’ it saved ‘high ranking military officer’ even though the sense was ‘in general’). This makes it hard to make sense of phrases and sentences within the article and is less effective in terms of acquisition of the lexis. I wonder whether it might be possible for the system to parse the phrase and make better translation suggestions, or better still, show a translation of the phrase and allow users to save chunks of language. And if this were too problematic from a translation point of view, some of the exercises could possibly focus on example sentences, rather than the word in isolation.
Sticking with the translation theme, there are also some more general issues with the quality of the material. For example, clicking on Spanish word ‘el’ saved ‘elevated railway’ to my inbox and I was once asked to match the word ‘gate’ to the translation ‘Bill Gates’! These dud translations can easily be removed for your inbox, but you still don’t know what ‘el’ or ‘gates’ means and there may be more subtle problems that users aren’t aware of when using the product.
Another drawback of the product is the lack of opportunity for output. The words are learned passively, with no need to produce or use communicatively. The Lingua.ly blog does have a lot of really useful advice for learners and lots of encouragement to speak, listen and communicate, but in the product itself the output and communicative aspects of language learning are lacking. It’s true that incorporating this in would quite drastically alter the product and maybe detract from the quality of what it does well, but from an SLA perspective, some opportunities for output would be helpful.
This lack of production also results in a lack of opportunities for personalisation and creativity within the product. It’s possible for users to add their own words manually and create heir own ‘word pack’ but this is only possible in the language that they’re learning and therefore words they already have some knowledge of. The exercise types are repetitive and therefore also without much chance of the learner bringing anything of themselves or their personality to the experience.