Depression and language: word analysis to identify suffering
In the project Fuori dal Blu (Out of the blue) ISTUD researchers pointed out that alone is one of the most recurrent words used by people diagnosed with major depression, revealing a condition of solitude experienced in solitude and not always understood by people around.
Several studies have questioned the relationship between language and depression. We present the work of Mohammed Al-Mosaiwi and Tom Johnstone, published on Clinical Psychological Science. An insight on language and depression, written by Al-Mosaiwi, is available on The Conversation.
The two authors say there are clear linguistic differences between people suffering from depression and other people. Within three different studies, Al-Mosaiwi and Johnstone conducted a textual analysis of 63 online forums, examining the use, from a linguistic point of view, of “absolutist” words (always, nothing, completely). According to the authors, anxiety, depression and suicidal ideation forums contain many more words of this type than others; moreover, they claim that “absolutists” words trace the seriousness of the disorders more faithfully than those describing negative emotions.
More specifically, the language used by people suffering from depression – again according to the authors – differs not only in words, but also in lexicon, grammar and length of sentences. For example, people who experience this condition, in addition to mostly using words recalling negative emotions (sad, lonely), often use first person singular pronoun, while they make little use of second and third person pronouns: according to the authors, this is a sign of how isolated these people are.
Understanding the language of depression can help us understand the way those with symptoms of depression think, but it also has practical implications. Researchers are combining automated text analysis with machine learning (computers that can learn from experience without being programmed) to classify a variety of mental health conditions from natural language text samples such as blog posts.
Such classification is already outperforming that made by trained therapists. Importantly, machine learning classification will only improve as more data is provided and more sophisticated algorithms are developed. This goes beyond looking at the broad patterns of absolutism, negativity and pronouns already discussed. Work has begun on using computers to accurately identify increasingly specific subcategories of mental health problems
The ultimate goal is trying to identify the symptoms of a depression or mental suffering in time: having more and more tools to identify such conditions is important to improve the quality of life and care of the affected people.