The Interplay of Science and Metaphor

Structured by AI 

Science, much like poetry, thrives on the use of metaphors. These figurative expressions can sometimes obscure meaning, leading to confusion or misinterpretation. At the heart of scientific inquiry lies the deployment of multiple metaphors, as exemplified by Pythagoras’s assertion that “All things are number.” This statement is often accompanied by another metaphor that helps frame the significance of those numbers, typically conveyed through models or tools.E. O. Wilson suggests that scientists should “think like poets and work like accountants.” This dual approach emphasizes that while precise calculations are essential, the true artistry of science lies in crafting effective metaphors. The ability to create compelling metaphors can elevate thinkers to genius status. For instance, Joule likened energy conservation to balancing a ledger, while Darwin drew parallels between natural selection and economic competition.However, the misuse of metaphors can lead entire disciplines astray. For example, equating people to “biological billiard balls” or economies to gases can result in significant misunderstandings. The economic poet Gary Becker’s metaphors—such as viewing families as “little firms” and children as “durable goods”—illustrate how poor metaphorical choices can distort reality.

The Dangers of Data-Driven Thinking

The mantra of letting “data do the talking,” popularized by proponents of Freakonomics, can be misleading. Alfred Marshall warned that relying solely on numerical data can be “treacherous.” Many concepts in fields like biology, economics, and social sciences—such as fitness, utility, and happiness—lack the measurable properties of physical quantities like mass or length. This limitation diminishes the effectiveness of mathematical approaches in these areas.There is often confusion regarding the relationship between quantitative and qualitative data. Nate Silver cautions that those who are not “quantitatively inclined” may inadvertently produce misleading conclusions. Effective quantification requires a solid foundation of qualitative understanding; otherwise, it risks generating nonsensical results. For instance, stating that the average human possesses one ovary and one testicle exemplifies the pitfalls of mixing different types of data.Statistical methods, while powerful, can be particularly slippery. They rely on the assumption that underlying phenomena exhibit stable patterns, which is often true for physical traits but not for behavioral data. This leads to logical fallacies, such as the fallacy of composition—where properties of parts are incorrectly assumed to apply to the whole—and its counterpart, the fallacy of division.

Missteps in Statistical Interpretation

Consider the contentious issue of police shootings. Sendhil Mullainathan’s assertion that racial bias in policing has “little effect” exemplifies the fallacy of division, as he assumes that national data accurately reflect local realities. Conversely, Rajiv Sethi highlights the fallacy of composition when questioning whether statistics from one city can be generalized to another with a different demographic makeup.Even leading researchers can mishandle statistical analysis, often engaging in practices like p-value cherry-picking or misapplying multiple regression techniques. Moreover, standard statistical methods may not always provide clarity; for example, randomization fails to address average testicle counts, and simply increasing data volume does not resolve inherent variability.

The Limitations of Quantification

Diane Coyle critiques GDP as a flawed measure, arguing that it fails to differentiate between harmful and beneficial economic activities and overlooks non-market contributions. The allure of data and calculation is strong, yet it is crucial to recognize that numbers do not hold exclusive rights to precision or truth. Words, metaphors, and qualitative insights can offer clarity and depth that numerical data sometimes cannot achieve.

Reference:

https://bigthink.com/hard-science/science-and-poetry-both-depend-on-metaphors/#link_time=1471092840