Protocol Sentences

Protocol sentences or protocol statements, also known as basic sentences or basic statements—the terms atomic statements, observation sentences, observation statements, and judgments of perception have also been used—were held by the logical positivists, verificationists, and empiricists to be the most elementary statements or sentences about empirical facts. They were the termination points of analysis of complex empirical sentences or statements into simpler components, or the basic elements out of which more complex empirical sentences or statements were supposed to be built or constructed. The epistemological view behind this held that the processes of verification or falsification or empirical investigation terminated in such protocol or basic sentences or statements.


A number of questions arose about such alleged protocol sentences: Do they actually exist? If they do exist, what relationship is there between them and non-verbal sensory (empirical) experience(s)? Are protocol sentences about private experiences of subjects, or is there anything public about them? If protocol sentences do exist, are they indubitable or incorrigible? Those questions were much discussed, especially by those who advocated the verifiability theory of meaning.

Wittgenstein's view

In the Tractatus Logico-philosophicus, Ludwig Wittgenstein held that a class of basic statements exists and that complexes are built up out of those elementary units. He held that an elementary proposition is an arrangement of names which represents a possible arrangement of simple things. It is—to use his metaphor or terminology—a picture of an elementary state of affairs.

Schlick's view

In Über das Fundament der Erkenntnis, Moritz Schlick questioned whether there exists a class of statements that provides an unshakable, indubitable foundation of all knowledge. He argued—against Carnap and Neurath—that such indubitable knowledge cannot arise from coherence with existing systems of science nor from someone's decision to accept a statement as true. He claimed that indubitable knowledge can be expressed only in a statement someone makes about his own immediate experience. Schlick called such statements Konstatierungen ("confirmations") and contrasted them with what Neurath and Carnap called "protocol sentences."

Schlick's Konstatierungen had a number of characteristics: Their form is "here, now, so and so," such as "here two black points coincide," or "here now pain." A Konstatierungen cannot be false because there is no difference between understanding it and verifying it. Unlike protocol sentences, Konstatierungen cannot be written down. They are not hypotheses. They are not the starting points of science, but are the means by which scientific hypotheses are confirmed.

One problem with Konstatierungen is that they are solipsistic. Another is that they cannot occur in predictions, since they are only reports of immediate experience. Thus, they cannot serve to confirm scientific hypotheses or theories because those embody predictions about future events. This means that they cannot serve a useful role in a philosophical account of scientific knowledge.

Carnap's view

The term "protocol sentence" itself is most closely related to the work of Rudolf Carnap. In his earlier work, Die physikalische Sprache als Universalsprache die Wissenschaft (1931), Carnap held that science is a system of statements built upon "primitive protocol sentences," which describe the experiences of scientific observers. They describe only what is directly given, so need no further justification.

Later on, in Logische Syntax der Sprache (1934), because of criticisms that had been given by Neurath, Carnap changed his view to the claim that what protocol sentences describe is not a factual but a linguistic question.

Neurath's view

In Soziologie im Physikalismus (1931/1932), Otto Neurath argued that sentences cannot be compared either with experiences of an observer nor with public material things, but only with each other. Some sentences, he held, are reports of direct observation acts. He also held that it must be possible to compare such protocol sentences with each other, and that this requires the existence of an inter-subjective language. According to Neurath, the form of a protocol sentence should contain the name of the observer plus a record of the act of observation. He gave as an example, "Otto's protocol at 3:17 o'clock [Otto's word-thought at 3:16: (In the room at 3:15 was a table perceived by Otto)]." The assumption here is that the sentence was written down by Otto at 3:17, reporting Otto's response at 3:16, recording Otto's perception at 3:15.

In Neurath's view the protocols of different observers, or of the same observer at different times, may conflict, and when that happens one or more of the protocols should be rejected; it is a matter of convenience and decision which to reject. The aim of science is to uphold a coherent system of statements (or sentences), but no one such statement is to be regarded as sacrosanct. Acceptance or rejection is, in the end, done on the basis of coherence and utility.

This view was strenuously rejected by Schlick, Russell, and Ayer because, they held, it distinguishes protocol sentences from others by purely syntactical methods; but, they held, a purely syntactical criterion of truth is wrong and is an abandonment of empiricism.

Russell's view

According to Bertrand Russell's account of knowledge by acquaintance and knowledge by description, every proposition that can be understood must be composed wholly of constituents with which people are acquainted. Sense data and universals are, Russell held, directly presented to the mind, so a person is acquainted with those. Although basic propositions are not incorrigible, the perceptual experience provides the strongest possible evidence for it. All empirical knowledge, he held, is built on such basic propositions.

The view of A.J. Ayer

A.J. Ayer considered at length whether basic propositions are indubitable or incorrigible. In "Basic Propositions" (1950), he held that if a sentence is a direct description of a private experience, then, although it may be verbally incorrect, it can not express a proposition about which the speaker can be wrong. In The Problem of Knowledge (1956) he argued that language rules may be private—a view that would be attacked by the later Wittgenstein in Philosophical Investigations. Ayer argued that such private language may be expressed in sense-datum terminology.

Popper's view

According to Karl Popper (In The Logic of Scientific Discovery), human experience cannot establish the truth of any statement. Instead, he held that people propose statements or claims and then test them. There does need to be a set of basic statements by which it can be decided whether a hypothesis or theory can be falsified, and a theory can be falsified by a basic statement only if the negation of that basic statement is derivable from the theory. Popper stipulated that the event referred to in the basic statement should be observable, but the expression "observable event" was an "undefined term which becomes sufficiently precise in use: …a primitive concept whose use the epistemologist has to learn…" (LSD). There are, he claimed, no grounds for stopping at any particular basic statement because all are subject to further tests if it is thought there is some need or reason to test them. It is a matter of agreement, not logic, where the testing process ceases.

To explain this, Popper used the metaphor of saying that human knowledge is like piles driven into a swamp. They are driven down until it is thought they are sturdy enough that edifice of knowledge can be erected upon them. If their sturdiness is unsure, they can be driven down more (by further testing basic knowledge claims). But bedrock can never be reached because there simply is no bedrock to be reached.

Rejections of basic statements

Later philosophers of science, especially Norwood Russell Hanson, Paul Feyerabend, and Thomas Kuhn, rejected the notion of any epistemological given because they held that there is no such thing as pure observation because observation is theory-laden. This means, as in an example given by Hanson, that the experienced investigator does not see the x-ray equipment and then conclude that the anode is overheating, but he sees it immediately as overheating. The inexperienced or unlearned observer, however, might not see anything at all in such a case. Popper made a similar point: He said that if a professor were to stand before a class and tell the students, "Observe, and write down what you observe," they could not fulfill the instruction unless they were told what to observe, or made some such further stipulation for themselves.

What this means is that the act of observing itself, including what is observed, depends on the knowledge and theoretic stance of the observer. And that undermines empiricism and positivism more-or-less thoroughly because empiricism and positivism depend on observation as the source of all knowledge. But if a conceptual or theoretical foundation—i.e. knowledge—must pre-exist in order that there can be meaningful observation, this refutes the empiricist claim that all knowledge must arise from perception.

Quine's reassertion of observation sentences

In a 1993 essay entitled, "In Praise Of Observation Sentences," W.V. Quine sought to restore an intermediate position between the received empiricist views on observation and the rejection of unmediated observation by Hanson, Feyerabend, Kuhn, et al. Quine claimed that, if one is "both single-mindedly physicalist and single-mindedly epistemological," that

Our channel of continuing information about the world is the impact of molecules and light rays on our sensory receptors; just this and some kinaesthetic incidentals. The protocol sentences should be the sentences most closely linked causally to this neural intake; most closely linked not in respect of subject matter, but physically, physiologically, neurally. They should be sentences like "It's cold," "It's raining," "That's milk," "That's a dog," to which we have learned to assent unreflectively on the spot if we are queried when certain associated sensory receptors are triggered (p. 108).

Quine went on to admit the objection that, "as vehicles of our evidence for our knowledge of the external world," such sentence "already assume such knowledge: Knowledge of rain, milk, dogs" (Ibid). But, he denied the objection on the grounds that they need not make that assumption:

Observation sentences—as I call them—can be conditioned outright to distinctive ranges of sensory intake, or as physicalists let us say neural intake. The child can be conditioned simply to assert or assent to the sentence under some distinctive stimulation, and he can come to learn only later that part of the sentence is a term denoting bodies or substances in an articulated external world (Ibid).

Quine wrote that "each of us learned some of his observation sentences in early childhood, and each of us learned most of them later." Some of them are learned by analogy, and people learn to form compounds of simple ones. "As adults," he claimed, "we learn many more through the mediation of sophisticated theory." So the chemist, for example, learns to use and understand the sentence, "There is some copper in the solution." Quine held that people learn the constituent parts of such a sentence "by construction from its separate parts," but "it becomes an observation sentence for a chemist who has learned to spot the presence of copper by a glance at a solution."

There are, Quine asserted, "degrees of observationality." Assertion or assent to an observation "may be more or less delayed or hesitant," and "There may even be afterthoughts" (p. 109), such as an observation: "Oh, it isn't a dog at all." So there is a self-correction stage, especially as the child becomes more sophisticated in making observation sentences and observation becomes infected by theory. But, Quine claimed, there are pure cases, "and they prevail at the early stages of language acquisition."

But this definitional account of observation sentences is, he wrote, only half of the story—the subjective and solipsistic half. The other half is "an intersubjective condition; for the sentences are learned from other speakers who are sharing the observation," and they need to be "intersubjectively tested." So an observation sentence "is Janus-faced. It faces outward to the corroborating witnesses and inward to the speaker. It faces outward to its subject matter and inward to the range of neural intake that is keyed to trigger it."

Quine conceded to those who have pointed out the theory-laden character of observation that "observation grades off into theory." But he tried to bridge over this problem by distinguishing "between taking observation sentences holophrastically and taking them piecemeal" (p. 110). "Piecemeal they are utterly theory-laden," he wrote, but considered holophrastically, "the observation sentences are anchored to sensory neural intake irrespective of their theoretical subject matter."

Quine held that this Janus-faced quality of observation sentences allows them to bridge the gap between observation language and theoretical language. "Theory is in logical contact with observation sentences through the sharing of terms," he wrote, because "[an] observation sentence faces neural intake when taken holophrastically and faces theory when taken term by term."

Whether or not Quine really solved the problems he intended to solve is open to question. He did admit that a word first acquired as "a fragment of a holophrastic observation sentence may undergo some semantic creep after being extracted from its observational context," and he gave the example of "whale" and "fish" or "weight" and "mass." But he dismissed this as merely "something the scientist adjusts for in relating his theory to his observations."

However, the whole notion of a "holophrastic observation" being beyond theory can be questioned. The online version of the American Heritage Dictionary defines the word "holophrastic" as "Of or relating to the stage of child language development characterized by the use of single-word utterances," and the Merriam-Webster Online Dictionary defines it as "expressing a complex of ideas in a single word or in a fixed phrase." Why should the child's one-word observation-sentence "Dog!" or "Pap!" not be considered to be just as theory-laden as the mechanic's observation on hearing a car engine and saying "Missing!" (meaning some cylinders not firing) or the photographer's lament on seeing a roll of negatives just pulled from the processor "Underexposed!" Both the child and the adult are making the observation based on already-learned theoretical considerations.

Quine was trying to preserve the foundations for an empiricism that would not fall victim to the criticism that it is not foundational but embedded in theory and theoretical considerations. But he failed at that, and the the fact that today's philosophers have mostly abandoned that attempt suggests that he and fellow empiricists were chasing a chimera.


  • Ashby, R.W. "Basic Statements." The Encyclopedia of Philosophy. New York: Macmillan, Inc., 1967.
  • Ayer, A.J. Language, Truth, and Logic. New York: Dover Publications, 1952.
  • Ayer, A.J. Foundations of Emprical Knowledge. London: Macmillan and Co., Ltd., 1951.
  • Ayer, A.J. "Basic Propositions," in Max Black, ed., Philosophical Analysis. Cornell New York: Cornell University Press, 1950.
  • Carnap, Rudolf. "Die physicikalishe Sprache als Universalsprache der Wissenschaft." Erkenntnis.Bristol, England: Thoemmes Press, 1995.
  • Carnap, Rudolf. Logische Syntax der Sprache. New York: Littlefield, Adams, 1959.
  • Hempel, Carl. "On the Logical Positivists' Theory of Truth." Analysis. Vol. 2, No. 4, 1934/1935.
  • Neurath, Otto. "Protokollsätze." Logical Positivism. Glencoe, IL: Free Press, 1959.
  • Neurath, Otto. "Soziologie im Physikalismus." Logical Positivism. Glencoe, IL: Free Press, 1959.
  • Popper, Karl. The Logic of Scientific Discovery. New York: Harper, 1968.
  • Popper, Karl. Conjectures and Refutations. London: Routledge and Kegan Paul, 1965.
  • Quine, W.V. "In Praise of Observation Sentences," The Journal of Philosophy. Vol. 90, No. 3. (Mar., 1993), pp. 107-116.
  • Quine, W.V.O. Methods of Logic. New York: Holt, 1959.
  • Quine, W.V. "Two Dogmas of Empiricism." In From a Logical Point of View. Cambridge, MA: Harvard University Press, 1953.
  • Quine, W.V. Word and Object. Cambridge, MA: Technology Press of the Massachusetts Institute of Technology, 1960.
  • Russell, Bertrand. The Analysis of Mind. London: G. Allen & Unwin, Ltd., 1921.
  • Russell, Bertrand. An Inquiry Into Meaning and Truth. London: G. Allen and Unwin, Ltd., 1948.
  • Schlick, Moritz. "Über das Fundament der Erkenntnis." Erkenntnis. Vol. 4, 1934.
  • Uebel, Thomas E. "Overcoming Logical Positivism from Within: The Emergence of Neurath's Naturalism in the Vienna Circle's Protocol Sentence Debate." Studien zur österreichischen Philosophie. Amsterdam: Rodopi, 1992. ISBN 905183442X
  • Wittgenstein, Ludwig. Tractatus Logico-philosophicus. New York: Harcourt, Brace & Company, Inc., 1922.
  • Wittgenstein, Ludwig. Philosophical Investigations. New York: Macmillan, 1953.

External links

All links retrieved June 11, 2015.

General philosophy sources


This article began as an original work prepared for New World Encyclopedia and is provided to the public according to the terms of the New World Encyclopedia:Creative Commons CC-by-sa 3.0 License (CC-by-sa), which may be used and disseminated with proper attribution. Any changes made to the original text since then create a derivative work which is also CC-by-sa licensed. To cite this article click here for a list of acceptable citing formats.

Note: Some restrictions may apply to use of individual images which are separately licensed.