Review | The Myth of Original Intentionality

Introduction

Long before my current immersion in complex systems, theoretical biology, and systems science—fields marked by self-organization, complexity, and emergence—I began my academic journey exploring what maybe many consider as the most complex architecture in our universe: the human brain. Such love by this organ was the result of being exposed early in my childhood to introductory psychology and education books, which my parents had at home because of their job as educators. Later in my youth I discovered the work of Daniel Dennett, whose book “Consciousness Explained”, inspired a myriad of doubts into my journey, catalyzing my intention to become a scientist.

A year ago, on April 19, Dan Dennett passed away. Today, on his first mournful anniversary, I revisit his seminal essay, “The Myth of Original Intentionality,” as a personal homage to a philosopher whose ideas profoundly influenced my intellectual trajectory. Selecting this essay from the diverse and inspiring Dennett’s bibliography was not an easy task. Such a manuscript was drawn from the book “Thinking Computers & Virtual Persons: Essays on the Intentionality of Machines,” which I discovered abandoned at my university. Despite having been published more than thirty years ago, I feel this piece is particularly apt amidst today’s fervent discussions surrounding Large Language Models and Artificial General Intelligence.

Intentionality and Artificial Intelligence

Dennett’s main objective in The Myth of Original Intentionality is to dismantle traditional conceptions of intentionality, primarily by challenging John Searle’s influential stance. Dennett meticulously differentiates between two closely related assertions:

• (S) Only an organic human brain—and certainly no electronic digital computer of the sort currently used in AI—could have the causal powers required to produce intentionality.
• (D) Only an organic human brain—and certainly no electronic digital computer of the sort currently used in AI—could have the causal powers required to produce the swift, intelligent sort of mental activity exhibited by normal human beings.

According to Dennett, proposition S succinctly encapsulates Searle’s argument, whereas proposition D aligns with his own perspective. He cautions readers against conflating these positions, despite their superficial similarity. Dennett argues that when Searle dismisses the causal efficacy of formal computer programs, this dismissal is trivially correct if one imagines a program as merely abstract symbols unimplemented in a physical system. Dennett emphasizes vividly that merely programming a non-mental object does not confer upon it the capacity for mental phenomena.

Dennett’s position initially appears paradoxical, given his well-known advocacy for artificial intelligence, framed largely within Church’s Thesis and a rigorous pursuit of clarity in psychological modeling. Personally, I find Dennett’s conflation between simulation and realization problematic, as genuine intelligence necessitates realizations grounded in adaptive semantics—an insight underscored by decades of limited progress when relying solely on symbolic, mere computational frameworks.

Despite these objections, Dennett maintains that his support for strong AI does not contradict proposition D. To him, authentic intelligence requires a richly interconnected neural system or its functional equivalent. He specifies two essential attributes necessary for intelligence: speed and massively parallel processing. Dennett underscores the role of speed, arguing it is indispensable for enabling appropriate organism-environment interactions, thereby implicitly highlighting the contemporary challenge that relevance poses—a problem increasingly recognized as resistant to purely algorithmic solutions. Additionally, Dennett asserts the indispensability of massively parallel processors, exemplified by the human brain, which enables rapid and complex cognitive interactions.

Original or Derived?

Central to Dennett’s argument is the concept of intentionality—fundamentally described as aboutness. Contrasting Searle’s claim of original intentionality as inherently mental, Dennett proposes a nuanced gradation of intentionality: real intentionality versus mere façon de parler intentionality. According to Dennett, intentionality attributed to artifacts or processes through interpretive stances—mentalistic projections of observers—is neither genuinely intrinsic nor truly derived but simply as if intentionality. This clarifies why the mysterious causal powers posited by Searle remain elusive; by definition, they fail to manifest through observable behavior.

Dennett argues compellingly that human beings themselves are artifacts shaped by evolutionary processes—survival machines engineered by genes. While genes do not possess intentionality in the literal sense, their evolutionary success relies on the intentionality inherent in natural selection, a slow yet precise process attuned to subtle rationales and relationships. Thus, human intentionality, characterized by foresight and self-representation, emerges as a highly specialized and derived outcome of evolutionary pressures.

In this way, Dennett resolves this apparent paradox by attributing genuine intentionality not to genes or individuals directly, but rather to natural selection itself. Although humans, robots, books, and maps exhibit derived intentionality, Dennett claims boldly that even human intentionality fundamentally derives from the intentional dynamics embedded within evolutionary processes. This leads him to a provocative conclusion: original intentionality simply does not exist.

Conclusion

Reflecting on Dennett’s provocative assertions on intentionality enriches our ongoing conversations about artificial intelligence, cognition, and the very nature of mind—highlighting his enduring impact on contemporary philosophy, science, and technology.

However, several gaps remain open, especially given contemporary advancements in plant neurobiology and AI. Dennett’s reliance on symbolic processing overlooks modern insights into embodied cognition and experimental evidence showing intentionality in living creatures which do not possess a central nervous system, which suggest that intelligence emerges from real-time interaction with the environment, rather than solely from symbolic manipulation. Moreover, current Large Language Models challenge Dennett’s claims about the necessity of parallel processing and speed by demonstrating the lack of cognitive capabilities despite using massively scaled sequential processing systems.

Dennett’s arguments also sidestep the problem of relevance, recently argued as inherently non-algorithmic, suggesting that computational modeling alone—no matter how sophisticated—may not fully replicate the nuanced intentionality and adaptability exhibited by living organisms. These contemporary insights urge a reevaluation of Dennett’s positions, prompting renewed discussions on the potential and limits of artificial intelligence in capturing the essence of intentionality and cognition.