Why Is Truth Never Universally Agreed Upon? Perspective and the Philosophy of Objectivity
Several winding paths cross a vast landscape and gradually converge toward the same distant horizon, suggesting that different perspectives can seek the same truth without becoming identical.
Why Is Truth Never Universally Agreed Upon? Perspective and the Philosophy of Objectivity
People disagree about truth despite sharing one reality. Explore how perspective shapes judgment and how objectivity can still be pursued.
Why Do People Interpret the Same Reality Differently?
People often expect the same evidence to produce the same conclusion. When disagreement persists, they may attribute it to ignorance, emotion, or error. Yet disagreement can arise even among informed and sincere observers because evidence does not interpret itself.
People may receive different information about the same event, give different weight to the evidence available, or rely on different background assumptions. Concepts, language, education, historical experience, and practical interests can influence which questions seem important and which explanations appear persuasive. These influences shape judgment without necessarily determining it completely.
Some disagreements concern facts, while others concern interpretation or value. Two historians may agree about when a revolution occurred but disagree about its causes or moral significance. People may accept the same scientific findings while judging their benefits and risks differently. A disagreement about meaning should therefore not automatically be treated as a disagreement about every underlying fact.
Several philosophical traditions, including hermeneutics and perspectivism, emphasize that human understanding always begins from a situated point of view. This does not prove that objective truth is impossible or that every interpretation is equally adequate. It shows why claims must be examined through evidence, reasoning, and criticism rather than accepted solely because they appear self-evident from one perspective.
Can Objective Truth Exist Beyond Perspective?
Truth, agreement, and objectivity are related but distinct concepts. A proposition does not become true merely because everyone accepts it, and it does not become false merely because no agreement has been reached. Universal agreement is therefore neither a necessary nor a sufficient condition for truth.
Philosophers have proposed competing accounts of what truth means. Correspondence theories connect truth with the way reality is. Coherence theories emphasize relations among beliefs or propositions. Pragmatist theories examine truth through inquiry and practice, while deflationary theories question whether truth names a substantial property at all. Philosophy has not reduced these approaches to a single accepted definition.
Objectivity also has more than one meaning. It can refer to a reality that exists independently of individual belief, or it can describe procedures designed to reduce the influence of personal preference and unexamined assumptions. Scientific inquiry usually pursues the second form through measurement, explicit methods, independent testing, replication, criticism, and revision.
These procedures do not automatically remove every bias. Experiments can be poorly designed, measurements can be incomplete, and peer review can fail. Objectivity becomes more credible when methods and evidence remain open to scrutiny and when qualified critics can challenge assumptions without being excluded.
Describing objectivity as a direction that inquiry approaches is therefore a useful metaphor, not a settled definition of truth. It expresses a regulative ideal: human observers remain limited, but disciplined inquiry can sometimes produce accounts that are less dependent on any one observer and more responsive to evidence.
Why Has Philosophy Never Stopped Questioning Truth?
In Plato’s dialogues, Socrates tests confident opinions by asking speakers to clarify their definitions and follow the implications of their claims. These dialogues often end without a complete resolution. Because Plato is the principal literary source for Socrates, however, the historical Socrates cannot always be separated with certainty from Plato’s philosophical presentation.
Descartes used methodical doubt for a different purpose. He suspended beliefs that could be doubted in order to identify a foundation he considered indubitable. His project was not to preserve doubt indefinitely but to reconstruct knowledge upon a secure basis.
Kant shifted the inquiry toward the conditions of human experience. He argued that space, time, and the categories of understanding help structure experience as human beings can know it. This does not mean that the mind arbitrarily creates reality. It means that knowledge of objects is conditioned by the cognitive capacities through which objects can become part of human experience.
Socrates, Descartes, and Kant did not share one theory of truth, nor did they agree that every certainty must remain permanently unsettled. What connects them more cautiously is the demand that important claims be supported by reasons rather than accepted solely through custom or authority.
Philosophical disagreements persist for many reasons: evidence may be incomplete, concepts may be ambiguous, methods and premises may differ, and new historical conditions may expose problems that earlier thinkers did not consider. Disagreement can reveal hidden assumptions and improve inquiry, but it can also result from error or confusion. It is intellectually valuable only when it remains answerable to reasons and evidence.
Can Artificial Intelligence Become an Objective Thinker?
Artificial intelligence systems can process large collections of data, compare arguments, and identify patterns that would be difficult for one person to examine. These abilities can support research and decision-making. They do not by themselves establish that an AI system understands truth or possesses an independent philosophical perspective.
AI outputs are shaped by more than training data. Model architecture, data selection, labeling practices, optimization goals, evaluation standards, prompts, and deployment settings all affect what a system produces. Human assumptions can therefore enter an AI system at several stages rather than through language and historical records alone.
AI can assist in identifying inconsistencies or patterns of unequal treatment, but it can also reproduce, obscure, or amplify existing errors. Its performance may vary across populations and situations, and fluent answers may contain unsupported claims. Claims about objectivity must therefore be tested through documented methods, appropriate evaluation, independent scrutiny, and continuing monitoring.
Responsibility for objectivity cannot be transferred to a machine. Researchers, institutions, developers, and users must decide which evidence is relevant, which errors are tolerable, whose interests are affected, and how decisions can be challenged. AI may contribute to this process, but it does not determine the standards of truth or justice by itself.
Truth is not created when perspectives reach agreement. Nevertheless, comparing perspectives can reveal blind spots that remain invisible within a single point of view. When those perspectives are tested against evidence, logical consistency, and a reality that can resist our expectations, disagreement can become a means of producing more adequate understanding.
References
Bender, Emily M., Timnit Gebru, Angelina McMillan-Major, and Margaret Mitchell. 2021. “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?” In ‘Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency’, 610–623. DOI: 10.1145/3442188.3445922.
Daston, Lorraine, and Peter Galison. 2007. ‘Objectivity’. Zone Books.
Descartes, René. 1985. ‘Meditations on First Philosophy’. In ‘The Philosophical Writings of Descartes’, Volume II, translated by John Cottingham, Robert Stoothoff, and Dugald Murdoch. Cambridge University Press.
Giere, Ronald N. 2006. ‘Scientific Perspectivism’. University of Chicago Press.
Kant, Immanuel. 1998. ‘Critique of Pure Reason’. Translated and edited by Paul Guyer and Allen W. Wood. Cambridge University Press.
Longino, Helen E. 1990. ‘Science as Social Knowledge: Values and Objectivity in Scientific Inquiry’. Princeton University Press.
National Institute of Standards and Technology. 2023. ‘Artificial Intelligence Risk Management Framework (AI RMF 1.0)’. NIST AI 100-1. DOI: 10.6028/NIST.AI.100-1.
Plato. 1997. ‘Complete Works’. Edited by John M. Cooper and D. S. Hutchinson. Hackett Publishing Company.
Tarski, Alfred. 1944. “The Semantic Conception of Truth and the Foundations of Semantics.” ‘Philosophy and Phenomenological Research’ 4 (3): 341–376. DOI: 10.2307/2102968.
굽이치는 여러 길이 광활한 대지 위를 가로지르며 하나의 먼 지평선을 향해 모여듭니다. 서로 다른 관점도 하나의 진실을 향해 나아갈 수 있음을 상징합니다.
진리는 왜 보편적으로 합의되지 않는가? 관점과 객관성의 철학
사람들은 하나의 현실을 공유하면서도 진리에 관해 의견을 달리합니다. 관점이 판단에 미치는 영향과 객관성을 계속 추구할 수 있는 이유를 살펴봅니다.
사람들은 왜 같은 현실을 다르게 해석할까?
사람들은 흔히 같은 증거를 접하면 같은 결론에 도달해야 한다고 생각합니다. 의견 차이가 계속되면 그 원인을 무지, 감정 또는 오류에서 찾기도 합니다. 그러나 증거가 스스로를 해석하는 것은 아니므로, 충분한 지식을 갖추고 성실하게 판단하는 관찰자들 사이에서도 의견 차이가 발생할 수 있습니다.
사람들은 같은 사건에 관해 서로 다른 정보를 얻거나, 주어진 증거에 서로 다른 중요성을 부여하거나, 서로 다른 배경 가정에 의존할 수 있습니다. 개념, 언어, 교육, 역사적 경험과 현실적인 이해관계는 어떤 질문을 중요하게 여기고 어떤 설명을 설득력 있게 받아들이는지에 영향을 줍니다. 그러나 이러한 요인이 판단에 영향을 준다고 해서 판단을 완전히 결정하는 것은 아닙니다.
어떤 의견 차이는 사실에 관한 것이고, 다른 의견 차이는 해석이나 가치에 관한 것입니다. 두 역사학자는 혁명이 일어난 시점에는 동의하면서도 그 원인이나 도덕적 의미를 다르게 판단할 수 있습니다. 사람들은 동일한 과학적 연구 결과를 받아들이면서도 그 이익과 위험을 다르게 평가할 수 있습니다. 따라서 의미에 관한 의견 차이를 모든 기초 사실에 관한 의견 차이로 간주해서는 안 됩니다.
해석학과 관점주의를 비롯한 여러 철학 전통은 인간의 이해가 언제나 특정한 위치와 관점에서 출발한다고 강조합니다. 그렇다고 해서 객관적 진리가 불가능하거나 모든 해석이 똑같이 타당하다는 결론이 입증되는 것은 아닙니다. 오히려 하나의 관점에서 자명하게 보인다는 이유만으로 주장을 받아들이지 않고, 증거와 추론과 비판을 통해 검토해야 하는 이유를 보여 줍니다.
관점을 넘어서는 객관적 진리가 존재할 수 있을까?
진리, 합의와 객관성은 서로 관련되어 있지만 동일한 개념은 아닙니다. 어떤 명제가 모든 사람에게 받아들여진다고 해서 참이 되는 것은 아니며, 합의에 이르지 못했다고 해서 거짓이 되는 것도 아닙니다. 따라서 보편적 합의는 진리의 필요조건도 충분조건도 아닙니다.
철학자들은 진리의 의미에 관해 서로 경쟁하는 여러 이론을 제시해 왔습니다. 대응론은 진리를 현실의 존재 방식과 연결합니다. 정합론은 믿음이나 명제 사이의 관계를 강조합니다. 실용주의 이론은 탐구와 실천을 통해 진리를 검토하며, 축소주의 이론은 진리가 실질적인 속성을 가리키는 개념인지 질문합니다. 철학은 이러한 입장들을 하나의 합의된 정의로 통합하지 못했습니다.
객관성에도 한 가지 의미만 있는 것은 아닙니다. 객관성은 개인의 믿음과 독립하여 존재하는 현실을 뜻할 수도 있고, 개인적 선호와 검토되지 않은 가정의 영향을 줄이기 위해 마련된 절차를 가리킬 수도 있습니다. 과학적 탐구는 일반적으로 측정, 명시적인 방법, 독립적인 검증, 재현, 비판과 수정을 통해 두 번째 의미의 객관성을 추구합니다.
이러한 절차가 모든 편향을 자동으로 제거하는 것은 아닙니다. 실험이 잘못 설계될 수 있고, 측정이 불완전할 수 있으며, 동료 평가도 실패할 수 있습니다. 방법과 증거가 공개적인 검토에 열려 있고, 자격을 갖춘 비판자들이 배제되지 않은 채 가정에 이의를 제기할 수 있을 때 객관성은 더 높은 신뢰를 얻습니다.
따라서 객관성을 탐구가 계속 접근하는 방향으로 설명하는 것은 진리에 관한 확정된 정의가 아니라 유용한 비유입니다. 이 비유는 하나의 규제적 이상을 나타냅니다. 인간 관찰자는 제한되어 있지만, 규율 있는 탐구를 통해 특정 관찰자에게 덜 의존하고 증거에 더 충실한 설명을 만들어 낼 수 있다는 뜻입니다.
철학은 왜 진리에 대한 질문을 멈추지 않았을까?
플라톤의 대화편에서 소크라테스는 자신 있게 제시된 의견을 검토하기 위해 대화 상대에게 개념을 분명하게 정의하고 주장의 결과를 끝까지 따져 보도록 요구합니다. 이러한 대화는 완전한 결론에 도달하지 않은 채 끝나는 경우가 많습니다. 그러나 소크라테스에 관한 주요 문헌이 플라톤의 저작이므로, 역사적 소크라테스와 플라톤이 문학적으로 제시한 철학적 인물을 언제나 확실하게 구분할 수 있는 것은 아닙니다.
데카르트는 이와 다른 목적을 위해 방법적 회의를 사용했습니다. 그는 의심할 수 있는 믿음을 잠정적으로 유보하고, 의심할 수 없다고 판단되는 토대를 찾으려 했습니다. 그의 목적은 회의를 영원히 유지하는 것이 아니라 안전한 기초 위에서 지식을 다시 세우는 것이었습니다.
칸트는 탐구의 초점을 인간 경험의 조건으로 옮겼습니다. 그는 공간과 시간, 그리고 오성의 범주가 인간이 인식할 수 있는 경험의 구성에 관여한다고 주장했습니다. 이것은 정신이 현실을 임의로 만들어 낸다는 뜻이 아닙니다. 대상에 관한 지식이 대상을 인간 경험의 일부로 받아들이게 하는 인지 능력의 조건을 받는다는 뜻입니다.
소크라테스와 데카르트와 칸트는 하나의 진리 이론을 공유하지 않았으며, 모든 확실성을 영구적으로 미결 상태에 두어야 한다는 데 동의하지도 않았습니다. 이들을 보다 신중하게 연결하는 요소는 중요한 주장을 관습이나 권위만으로 받아들이지 않고 근거를 통해 뒷받침해야 한다는 요구입니다.
철학적 의견 차이가 지속되는 데에는 여러 이유가 있습니다. 증거가 불완전할 수 있고, 개념이 모호할 수 있으며, 방법과 전제가 다를 수 있고, 새로운 역사적 조건이 이전 사상가들이 고려하지 않았던 문제를 드러낼 수도 있습니다. 의견 차이는 숨겨진 가정을 밝히고 탐구를 발전시킬 수 있지만, 오류나 혼란에서 생길 수도 있습니다. 의견 차이는 근거와 증거에 응답할 때에만 지적인 가치를 지닙니다.
인공지능은 객관적인 사유자가 될 수 있을까?
인공지능 시스템은 방대한 자료를 처리하고, 여러 주장을 비교하며, 한 사람이 검토하기 어려운 패턴을 찾아낼 수 있습니다. 이러한 능력은 연구와 의사결정을 지원할 수 있습니다. 그러나 그 능력만으로 AI가 진리를 이해하거나 독립된 철학적 관점을 지닌다고 입증되는 것은 아닙니다.
AI의 출력은 학습 자료에 의해서만 형성되지 않습니다. 모델 구조, 자료 선별, 분류 기준, 최적화 목표, 평가 기준, 사용자 입력과 실제 사용 환경이 모두 결과에 영향을 줍니다. 따라서 인간의 가정은 언어나 역사 기록뿐 아니라 AI 시스템을 개발하고 사용하는 여러 단계에서 개입할 수 있습니다.
AI는 불일치나 불평등한 처리 양상을 발견하는 데 도움을 줄 수 있지만, 기존의 오류를 재현하거나 감추거나 확대할 수도 있습니다. 집단과 상황에 따라 성능이 달라질 수 있으며, 유창한 답변에도 근거 없는 주장이 포함될 수 있습니다. 따라서 객관성에 관한 주장은 문서화된 방법, 적절한 평가, 독립적인 검토와 지속적인 관찰을 통해 검증해야 합니다.
객관성에 대한 책임을 기계에 이전할 수는 없습니다. 연구자, 기관, 개발자와 사용자는 어떤 증거가 중요하고, 어떤 오류를 허용할 수 있으며, 누구의 이해관계가 영향을 받고, 결정에 어떻게 이의를 제기할 수 있는지 판단해야 합니다. AI는 이 과정에 기여할 수 있지만, 그 자체로 진리나 정의의 기준을 결정하지는 않습니다.
관점들이 합의한다고 해서 진리가 만들어지는 것은 아닙니다. 그러나 서로 다른 관점을 비교하면 하나의 관점 안에서는 보이지 않던 맹점을 발견할 수 있습니다. 그러한 관점들이 증거와 논리적 일관성, 그리고 인간의 기대에 저항할 수 있는 현실을 기준으로 검토될 때 의견 차이는 더욱 적절한 이해를 만들어 내는 수단이 될 수 있습니다.
참고문헌
벤더, 에밀리 M., 팀닛 게브루, 안젤리나 맥밀런메이저, 마거릿 미첼. 2021. 「확률적 앵무새의 위험에 관하여: 언어 모델은 지나치게 거대해질 수 있는가?」 ‘2021년 ACM 공정성·책임성·투명성 학술대회 논문집’, 610–623쪽. DOI: 10.1145/3442188.3445922. 대규모 언어 모델의 출력이 학습 자료와 개발 조건의 영향을 받으며, 유창성이 이해나 진실성을 보장하지 않는다는 점을 검토한 논문입니다.
대스턴, 로레인, 피터 갤리슨. 2007. ‘객관성’. 존 북스. 과학에서 객관성이라는 이상이 역사적으로 어떻게 형성되고 변화했는지를 분석한 연구서입니다.
데카르트, 르네. 1985. ‘제일철학에 관한 성찰’. 존 코팅엄, 로버트 스투트호프, 두걸드 머독 옮김, ‘데카르트 철학 저작집’ 제2권 수록. 케임브리지대학교출판부. 방법적 회의가 회의 자체의 유지가 아니라 확실한 지식의 토대를 찾기 위한 절차였음을 확인할 수 있는 원전입니다.
기어리, 로널드 N. 2006. ‘과학적 관점주의’. 시카고대학교출판부. 과학적 지식과 모델이 특정한 관점에서 현실을 나타내면서도 객관적인 지식을 제공할 수 있는 조건을 논합니다.
칸트, 임마누엘. 1998. ‘순수이성비판’. 폴 가이어, 앨런 W. 우드 옮김·편집. 케임브리지대학교출판부. 인간의 경험과 인식이 감성과 오성의 형식에 의해 조건 지어진다는 칸트의 논의를 확인할 수 있는 원전입니다.
롱기노, 헬렌 E. 1990. ‘사회적 지식으로서의 과학: 과학적 탐구의 가치와 객관성’. 프린스턴대학교출판부. 과학적 객관성이 개인의 중립성만이 아니라 공개적인 비판과 사회적 검토 과정에 의존한다는 점을 논증합니다.
미국 국립표준기술연구소. 2023. ‘인공지능 위험관리 프레임워크 1.0’. NIST AI 100-1. DOI: 10.6028/NIST.AI.100-1. AI의 타당성, 신뢰성, 편향, 투명성과 책임성을 지속적으로 평가해야 한다는 공인 지침입니다.
플라톤. 1997. ‘플라톤 전집’. 존 M. 쿠퍼, D. S. 허친슨 편집. 해킷출판사. 소크라테스의 문답과 철학적 검토 방식을 전하는 주요 원전들을 수록하고 있습니다.
타르스키, 알프레트. 1944. 「진리의 의미론적 개념과 의미론의 기초」 ‘철학 및 현상학 연구’ 4권 3호, 341–376쪽. DOI: 10.2307/2102968. 형식 언어에서 진리를 엄밀하게 정의하려 한 대표적인 현대 논문입니다.
Previous Episode - How Does Doubt Become Knowledge? From Skepticism to the Scientific Method
Related Label - Philosophy
Next Episode - Plato's Republic — How Does Truth Lead a Community?

Comments
Post a Comment