Interpreting the Endangered Languages from the Linguistic Relativity Perspective and Utilizing Machine Translation as a Preventive Measure

Interpreting the Endangered Languages from the Linguistic Relativity Perspective and Utilizing Machine Translation as a Preventive Measure

PRELUDE

    “The limits of my language mean the limits of my world.”, Ludwig Wittgenstein said… For centuries, philosophers have engaged in the relationship between the language and its significant impact on individuals and societies which is leading to the point that the potential loss of a language may be carrying severe implications for cultural diversity. In that propose, could we preserve languages through using technological advances such as machine learning techniques and revive them again?

REPRESENTATION IN MEDIA

    The way how language and human behaviour correlate or the ultimate relationship between these terms might be difficult to comprehend at first. Moreover that can be interpreted in two distinct ways as, each offers different aspects of languages’ impact on we human beings’ behaviours and cognitive processes in general. Let’s begin by discussing the concept we can refer to as the rhetorical standpoint. That concept essentially offers to utilize the language as a tool for persuasion. “Friends, Romans, countrymen, lend me your ears” the famous quote from William Shakespeare’s Julius Caesar might not only serve as a prominent example of the concepts representation in media, but also how effective of its utilization might be by military leaders and statesmen from ancient times up to the modern age. As by recalling these lines which Mark Antony succeeds to influence the crowd can attest to its strong impact as a rhetorical strategy. Yet, what we focus in this article is more of a behaviorist theory aspect within the context of language. I believe that this point of view could provide us more insights for the rest of the article.

    “If thoughts can corrupt language, language can also corrupt thoughts.” In George Orwell’s well-known work 1984; the government introduces Newspeak, a language, aimed to gain influence over people to execute its own agenda. By restricting vocabulary and altering linguistic structures, the government seeks to limit people’s awareness and limit them to make a move against government’s inconsistent actions. Eventually these tactics perform well that people lose the ability to be sensitive towards government’s actions as they can not even relate to such concepts that could evolve an opposition movement against the government. Another reference in this manner can be drawn to the language, spoken by the people who live in the fictional world of Anarres in Ursula Le Guin’s The Dispossessed . In Pravic, there are no such things as possessive pronouns. This feature of the language reflects the people of Anarres’ collective ownership based lifestyle over individual ownership. If there is no such concept as individual ownership, there’s also no point in using possessive pronouns in the language, right? Perhaps they might have thought like this… Before we take a deep dive into the examination of endangered languages’ aspect, I believe, it would be helpful to revisit the ancient times to meet the roots of linguistic relativity which we already covered the media representations and grasped the essential concept of.

LINGUISTIC RELATIVITY ASPECT

    In Plato’s Cratylus dialogue, profound notions can be seen regarding the connection between language and the essence of reality. Through the dialogue with Cratylus, he advocates for the notion that words should ideally reflect the true nature of the objects, they signify in linguistic context. In Plato’s discourse on the interrelation between language and reality within the dialogue, he recognizes the crucial role of language in both understanding and shaping reality. Yet, he also underscores the inherent limitations of language; thus, the word could only provide imperfect representations of the true aspects of reality. Well at this point, if Plato’s assertion doesn’t make sense at all, let’s imagine an apple. You probably tend to imagine a shiny red apple with no rotten spots but for another person, that imanigation could include a Granny Smith apple with rotten spots. However in the real world there are a bunch of apples you can see that don’t fit either of them. From the point we comprehend Plato’s notion that underscores the language and its imperfect representations of reality, what if we come up with “how big are the chances of any major differences between two different people’s representations of reality based on the language they speak”? Apparently, before the Romanticism age, philosophers weren’t interested in as there was little scholarly attention directly given to the relationship between individual languages and its effects on different thought patterns. Nevertheless following the French Revolution, nationalism surged in popularity; leading to increased focus on shared cultural values among the nation, brought attention to the languages and their impacts on nations. The principle of linguistic relativity, commonly referred to as the Sapir-Whorf hypothesis, is attributed to linguists Edward Sapir and Benjamin Lee Whorf, who played major roles in advancing its prominence within academic discourse. Their contributions were notably influenced by the insights of Alexander von Humboldt. According to Humboldt, each language presents a unique worldview, influencing the cognitive structures of its speakers[1]. A shared linguistic context proceeds commonalities in the experiences and perceptions of the world among individuals who speak the same language. This concept developed into linguistic determinism apart from its initial version, proposing that language and its structures constrain and shape individual thought processes. This notion represents a strong interpretation of linguistic relativity, covers of a concept depicts language as having an ultimate impact on cognition. Another movie can be referenced in the context of this proposition, Arrival, where a linguist is recruited to establish communication with extraterrestrial visitors. Through the process of linguistic engagement, the protagonist acquires the capacity to perceive temporal dimensions beyond conventional human cognition, as she progressively masters the alien language. Even though that can be seen as an instance for linguistic determinism as well as the references above, this one stands out from prior instances as it directly exemplifies the concept that language can fundamentally alter our comprehension of reality and perception. With that being said, no matter how fun it sounds like to believe someone becomes able to adapt such talent just by learning another language, linguistic determinism lacks widespread recognition among the academic community as you might have expected.

UTILIZING MACHINE TRANSLATION AND OTHER ML-BASED TECHNOLOGIES

    Considering Plato’s point on the inherent imperfections of language and Sapir-Whorf hypothesis may lead us to think that languages convey all these imperfect ways of interpretations towards the reality, what will be the consequences of a language loss. From my perspective, it can be also interpreted as a loss of unique perspective a within the broader scale. In Te reo Māori, the Indigenous language of Aotearoa New Zealand, distinct conceptualizations of the world exist by the words such as “te taha wairua” (spiritual world) and “te taha kikokiko” (physical world). Unfortunately, the vulnerable status of over 1500 endangered languages, each with its own unique terms and perceptions as the language of Aotearoa has, at the close risk of extinction by the century’s end. This situation can be addressed in a few ways, including raising awareness and and actively promoting the preservation of endangered languages. Google launched a machine learning-based application aimed at providing translations of objects into endangered languages Woolaroo. Google anticipates that this endeavor will contribute to raising awareness and proceeding greater interest in the preservation of these languages. The introduction of the project as a “pilot experiment” implies the prospect of its further progression and evolution. Apart from Woolaroo and the several solutions by the other initiatives on raising awareness and promoting languages; advancement of neural machine translation techniques allows us to achieve higher quality translations than ever before, even with the limited data sets. In 2023, a scholarly article was published where a group of researchers detailed their success in translating Akkadian into English, using Neural Machine Translation (NMT). That advancement offers a hope for utilizing the same techniques to preserve the other endangered languages towards the risk of information loss. It seems also possible to create more texts and content for endangered languages thanks to Large Language Models (LLM). They are used for text generation powered by artificial neural networks. Following Google’s introduction of the Transformer neural network architecture in 2017, featuring innovations such as Positional Encoding and Self-Attention, artificial neural network models underwent a significant advancement. This innovation enabled developers to train models more efficiently with limited data and surpassing the previous technique known as “Recurrent Neural Networks” (RNNs). GPT-1 (Generative Pre-trained Transformer 1) was developed based on Google’s architecture transformer by OpenAI one year subsequent to Google’s innovation. As of April 2024, GPT-4 has been introduced competing alongside Google’s Gemini 1.5 and Claude 3 by Anthropic. Each model shows remarkable performances on generating texts in even low resource languages and the ongoing progress of these models suggests promising implications in terms of creating more content to revitalize these endangered languages.

SUMMARY

    As the time passes, the significance of what we are losing becomes increasingly apparent. Among these losses are languages, each holding the thoughts, feelings, heritage, and history of its people. Whether languages directly impact individuals’ minds or not, preserving them remains crucial. As ML and AI-based systems continue to evolve and revolutionize numerous aspects of our lives, the latest technological advancements inspire us with optimism in this manner. After all is said and done, if they are going to bring a cure for endangered languages or how significant the impact will be, presumably, time will tell…

References:

  • Doğru, S. G., & Dombayci, M. A. (2023). WORLD AND LANGUAGE: PHILOSOPHICAL ASSESSMENT OF LINGUISTIC
    RELATIVITY THOUGHT. FLSF Felsefe Ve Sosyal Bilimler Dergisi, 35, 87–112. https://doi.org/10.53844/flsf.1244176
  • Humboldt, W. von. (1836). On Language.
  • Le Guin, U. K. (1974). The Dispossessed. New York: Harper & Row.
  • Orwell, G. (1949). 1984. London: Secker & Warburg.
  • Wittgenstein, L. (1921). Tractatus Logico-Philosophicus. London: Kegan Paul.

Visual References:

Author

0 0 votes
Article Rating
Subscribe
Notify of
guest

0 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
0
Would love your thoughts, please comment.x
()
x