• Ocean

From Hippocrates to the Big Five and AI: A History of Personality Models

Who am I? The perennial question that the “Big Five” personality model attempts to answer. Originating in the 1980s, the model was a significant scientific breakthrough and remains the most popular and supported theory for understanding personality today. In this blog post, I’ll look closer at the model’s history, its current state and relevance to DataSine.

Our personalities are shaped by a variety of environmental and biological factors and affect our emotions, thoughts and behaviour. They are unique to each person and persistent through time. Most modern psychologists believe that personality has five basic dimensions, often referred to as the “Big Five”. They can be represented with the acronym – OCEAN, which stands for Openness, Conscientiousness, Extraversion, Agreeableness and Neuroticism.

Throughout the years, many theories have been put forward, arguing for varying numbers of different personality traits – Gordon Allport found 4,000, Raymond Cattell 16, and Hans Eysenck three. But with the Big Five taxonomy, psychologists finally managed to approach a consensus on personality dimensions. Currently, it is the most respected, widely studied method for characterising an individual’s personality. But how did it begin?

Early Models

While there are many differences, all personality models –  from the Four Temperaments of Ancient Greece to the modern day Big Five –  are underpinned by the idea that we can understand ourselves and other people by analysing how we experience, respond and act towards our physical and social environment.


Early personality models were based on observable behaviours. The earliest known is the Four Temperaments from Hippocrates around 400 BC. The model identifies four personality types:

  • sanguine: someone who is enthusiastic, active and social – this share similarities with Extravert trait from the Big Five
  • choleric: someone who is independent, decisive and goal orientated
  • melancholic: someone who is analytical, detail oriented and deep thinkers.
  • phlegmatic: someone who is relaxed, peaceful and quiet – this shares similarities with the Introvert trait from the Big Five.

According to Hippocrates, personality was influenced by levels of four bodily fluids:  blood, yellow bile, black bile, and phlegm. Around the same time, Plato proposed four slightly different temperaments: – iconic (i.e. artistic), pistic (i.e. sensible), noetic (i.e. intuitive) and dianoetic (i.e. reasoning).

Modern Day

You can split modern-day personality research into two broad strands: research that is underpinned by Cognitive Theory and research that is underpinned by the Lexical Hypothesis. The former seeks to explain personality traits according to how we take in and process information while the latter argues that all personality traits are encoded in natural language. As such, models based on the Lexical Hypothesis are built from studying actual data whereas models based on Cognitive Theory are built from theories about what goes on in people’s heads.

Cognitive Theory: Top-Down Approach

In 1932 Sigmund Freud – the father of psychoanalysis – published The Ego and the Id. Freud believed that personality had three aspects: the id, ego and superego, all developing at different stages in our lives. The id is the primitive and instinctual part of the mind that holds sexual and aggressive drives and hidden memories, the super-ego works as a moral conscience, and the ego is the realistic part that mediates between the id and the super-ego. Interacting to form a whole, they contribute to a person’s behaviour. Although a very interesting theory philosophically, it was hard to verify scientifically.

Carl Jung, a psychiatrist and student of Freud, claimed there were eight different types of consciousness. In one’s attitude towards the world, a person could either be extraverted or introverted. In order to access and process information about the world, a person could four different functions: feeling, thinking, sensation and intuition. Jung’s ideas were further popularised by Katherine Cook Briggs and her daughter Isabel Briggs Myers who incorporated these categories in their personality test –  the Myers-Briggs Type Indicator (MBTI). Although not scientifically validated and thoroughly researched, MBTI remains hitherto popular.

The Lexical Hypothesis: Bottom-Up Approach

Data-driven approaches to understanding personality arguably began with Sir Francis Galton in 1884. The sociologist was the first to put forward the Lexical Hypothesis, which argued that personality becomes encoded in language over time and by analysing language, you can establish what personality traits exist. His approach was further developed in the 1930s by Gordon Allport and Henry Odbert at the National Institutes of Health in the US. They examined Webster’s New International Dictionary and identified over 17,000 words that described personality or behaviour.

In the 1940s, psychologist Raymond Cattell applied factor analysis – a statistical technique for identifying trends in large data sets – to lists of words that describe personality characteristics. He found that 16 factors accounted for the variance in personality descriptions. Two decades later, Ernest Tupes, Raymond Christal and Warren Norman condensed these into five factors: surgency, agreeableness, conscientiousness, emotional stability and culture.

Separately, Lewis Goldberg conducted research based on the lexical hypothesis, also using factor analysis, and found five principal dimensions to personality that were very similar to those identified by Tupes, Christal and Norman: openness, conscientiousness, extraversion, agreeableness and neuroticism. These became the “Big Five” we know today.

The “Big Five”, AI and DataSine

In 1993 Oliver P. John, Eileen M. Donahue and Rachel L. Kendal constructed the Big Five Inventory (BFI) as a clear, simple and efficient way to measure an individual’s Big Five personality. The BFI focused on the trait adjectives, using short phrases to measure each Big Five Dimension. For example, the dimension of Openness has an adjective of ‘original’, which became the BFI item ‘“Is original, comes up with new ideas”.

In 2016 Oliver P. John, in collaboration with Christopher J. Soto – who we now work closely with at DataSine – developed a major revision of the BFI with the Big Five Inventory-2 (BFI-2).  This revision resulted in a personality measurement tool that provides greater reliability, breadth, accuracy and predictive power. BFI-2 is an important tool for us here at DataSine.

Technological advances and huge increases in available data have been game changers for personality research. It’s these changes that we are embracing at DataSine – using advanced machine learning to not only predict personality but provide recommendations for what content is likely to appeal most to different personality traits. As the process is automated, understanding people’s personality and personalising content accordingly becomes possible at scale. The result? Organisations benefit from increased customer engagement, loyalty, and retention; customers feel more valued and enjoy an improved user experience.