The Hedgehog and the Fox

Recently, while reading ‘On Grand Strategy’ by John Lewis Gaddis, I came across the philosopher Isaiah Berlin’s work. Gaddis’ use of the classification framework of foxes and hedgehogs based on Isaiah Berlin’s 1953 book The Hedgehog and the Fox: An Essay on Tolstoy’s View of History struck me as particularly interesting. In his book, Berlin builds on a fragment attributed to Greek philosopher Archilochus, who supposedly stated that “a fox knows many things, but a hedgehog one important thing.” Expanding upon this quote, Berlin suggests that writers and thinkers could potentially be classified as either foxes or hedgehogs. Foxes, according to Berlin, have a wide variety of interests and rely on a great number of experiences and sources for their decision-making. While expounding on the traits of foxes in greater detail, Berlin names Aristotle and Goethe as potential examples of thinkers with fox-like characteristics. Hedgehogs, on the other hand, tend to base their interpretation of the world on a single, defining idea. To illustrate his point, Berlin lists thinkers such as Plato and Nietzsche as thinkers with hedgehog-like traits.

Data Science Projects in Big Organizations

While at first glance, Berlin’s framework may offer relatively little value for data science projects, examining the core idea behind it reveals an entirely different picture. Due to the recent boost in the popularity of data science, many organizations that are not primarily invested in the technology industry have started hiring data scientists. However, the circumstances that these data scientists find themselves in are profoundly different from those in tech companies. For one, the supervisor in charge of the data science team might very well not have a data science or analytics background at all. As a result, the data science team will have a much harder time communicating its proposed projects to management. Besides the people in charge of the budget not fully understanding the data science teams’ needs, the data scientists might also find themselves confronted with a significant amount of skepticism towards their projects from the general workforce of large organizations. Whilst there could be many drivers behind this skepticism, the most prevalent ones generally stem from a lack of understanding of what the data science team actually does. Organizations whose business has little to do with tech generally have a workforce that perhaps knows how to utilize Microsoft Office (if you are lucky). Thus, the knowledge gap between data scientists and the remaining workforce is quite significant from the get-go. Adding to this confusion, in my opinion, is labeling everything as ‘AI’. Most people, when confronted with the term ‘Artificial Intelligence’ immediately think of robots that have the same capabilities as human beings or other science fiction figments. In doing so, people build up a natural hesitance towards everything that the data science team does, assuming it to be too complicated to understand and thus not worthy of investing their time into it in the first place.

Should a Data Scientist Be a Hedgehog or a Fox?

The classification framework proposed by Berlin relates to the challenges data science teams find themselves confronted with in several ways. In the following paragraphs, I will attempt to extend Berlin’s framework to data scientists in large organizations by applying the general concepts to specific challenges in the daily life of data scientists.

Let us first examine what characteristics might lead to a data scientist being classified as a hedgehog. In my opinion, a data scientist could qualify as a hedgehog if he or she exhibits an unrelenting focus on their own motivations. When proposing a project, he or she is not concerned with the opinions and worries of non-technical stakeholders (who might actually be the ones in charge of the budget) but only with her own fascination with this particular project. While there is absolutely nothing wrong with being captivated by the technical aspects of a project, quite the contrary, trying to pursue data science projects using this mindset will cause tremendous problems in non-tech organizations. By only being able to view the project from the point of view of a data scientist, you will very quickly lose sight of the thoughts and concerns of people in charge of the budget. In communicating with management, especially at non-tech organizations, project approval is highly correlated with management being able to see how your project creates value for the organization. Value, in most cases, is either a cost reduction or an increase in revenue generated. Therefore, much of the difficulties that a data scientist hedgehog encounters directly relate to the ability to approach issues from various angles and being capable of framing your project proposals accordingly.

In contrast to hedgehogs, foxes rely on a plentitude of inputs to help guide them. Intuitively, this seems like the more beneficial of the two relating to data science after having established that a narrow focus on technical aspects can bring about various issues. Nevertheless, only being a fox in data science does not necessarily have to be better than only being a hedgehog. While strictly focusing on your point of view can be harmful to getting support for your project, trying to incorporate everybody else’s opinions into your project can doom the project’s success just as much. Assuring everyone around you that you will adapt your project according to their needs might simplify securing funding for the project, however, it might contort your project to the point where it is an entirely different project. In my opinion, the issue of having to pitch a project to someone not familiar with data science at all is no best solved by almost completely giving up on your input and instead purely relying on the opinions of people that do not have as deep of an understanding of the matter at hand as you do. In that case, I think being able to stand your ground when it comes to the core concept of the project is of great importance.

Conclusion

All in all, it seems to me that having the ability of a hedgehog to maintain focus on your final goal is crucial in order to get your data science project done. At the same time, not getting caught up in your personal perceptions and way of thinking but instead trying to understand where other stakeholders are coming from and how to best approach them is equally important to me. I am not sure whether this conclusion fits the framework proposed by Berlin, however, Berlin himself is reported to have never wanted people to take this classification too seriously. Nonetheless, the hedgehog and fox framework can provide a stimulus for an important discussion.

If you have any thoughts on the article, feel free to reach out to me. I would be happy to hear and discuss your opinions.

References:

[1] Gaddis, Joh Lewis, On Grand Strategy (2018), Penguin Press

[2] Berlin, Isaiah, The Hedgehog and the Fox (1953), Weidenfeld & Nicolson