Sense-Data and Adverbialism

Sense-data theory and adverbialism accept a common internal element across veridical perception, illusion, and hallucination. They disagree about its ontology. A sense-datum theorist says the subject is directly aware of a private object bearing the apparent qualities. An adverbialist says no object is required: experiencing is an event or act modified in a redly, circularly, or painfully manner.

Both were developed to preserve the evident fact that an experience can occur when its ordinary external object does not. Sense-data supply a bearer for experienced qualities; adverbialism avoids multiplying objects. Their respective problems - a veil between mind and world, and difficulty representing complex structure - helped motivate intentionalism and disjunctivism.

What is a sense-datum?

A sense-datum is the immediate object of perceptual awareness. It is not the retinal image or a neural state, since those are not normally seen by the subject. It is a mind-dependent item whose properties correspond to how things look: a red, round patch in visual experience, for example.

In a veridical case, an ordinary object causes an appropriate sense-datum. In illusion, the datum has a property the object lacks. In hallucination, a similar datum occurs without the ordinary object. The theory provides a stable common factor and respects the thought that awareness must be awareness of something.

Sense-data can be understood as particulars, property instances, or contents. The more they are weakened into mere representational contents, the closer the theory moves to intentionalism and away from the distinctive claim that a private object is perceived.

The phenomenal principle

The argument depends on a phenomenal principle:

If it sensibly appears to a subject that something has qualitative property $F$, then the subject is aware of something that actually has $F$.

When a white wall looks yellow, something yellow must be present to awareness. Since the wall is not yellow, a sense-datum is.

The principle is not self-evident. "The coin looks elliptical" need not mean an elliptical item is seen; it may describe the way a circular coin appears from an angle. Perceptual predicates can be non-committal: "looks $F$" reports a presentation without quantifying over an $F$ object.

Rejecting the principle allows representation or adverbial modification to carry the apparent property. Defending it requires an argument that qualitative awareness always needs an instantiated bearer.

The argument from hallucination

Hallucination strengthens the case because no ordinary bearer exists. If a subject is aware of a red circle when no red circle is present, an internal red circular item seems to supply the object.

The move assumes that hallucination is genuinely relational: awareness must terminate in an object. An intentionalist instead says the state represents a red circle without one existing, just as belief can concern Pegasus. An adverbialist says the subject senses redly and circularly. A disjunctivist may deny that hallucination has the same positive nature as perception.

Sense-data solve the missing-object problem by adding an object. The alternatives solve it by rejecting the relational assumption in bad cases.

The veil of perception

If subjects directly perceive only sense-data, ordinary objects stand behind a veil of perception. Evidence for the external world consists of private items that could, in principle, occur without their normal causes. Sceptical scenarios become difficult to exclude.

The sense-datum theorist can appeal to inference to the best explanation: stable, coherent sense-data are best explained by external objects. But perception then seems less direct than ordinary knowledge assumes. A subject knows a tree through an inference from an inner object, even when no conscious inference occurs.

Defenders may distinguish epistemic mediation from conscious reasoning. A reliable causal relation can provide knowledge without explicit inference. The remaining concern is why the private intermediary is needed if neural processing can realise direct representation of the external object.

Duplication and location

Sense-data appear to duplicate qualities. A red apple causes a red datum; the theory must explain the relation between physical surface reflectance, perceptual processing, and private redness. If the datum is literally coloured, where is it, and under what illumination does it possess colour?

Placing the item in mental space avoids physical location but introduces a second kind of spatial structure. Treating colour as a mental property avoids category mistakes but makes the resemblance between datum and object metaphorical.

The ontological cost would be justified if sense-data did explanatory work unavailable to representational states. Their critics argue that the theory turns features of an act or content into additional objects without improving prediction.

The act-object ambiguity

Experience-talk can grammatically suggest an object. "I have a red visual impression" may refer to an event of being visually impressed, not to a red thing possessed by the subject. The act-object ambiguity supports adverbialism.

Compare "she danced gracefully." Gracefulness modifies the dancing; it does not name a graceful object that she encountered. Likewise:

$$\text{experiences a red circular datum}$$

can be reconstructed as

$$\text{senses redly-and-circularly}.$$

The subject and experiential event exist, but no object bearing the presented qualities is added.

Adverbialism

Adverbialism treats experiential qualities as manners of sensing. Hallucination poses no missing-object problem because sensing redly does not entail that anything red exists. Veridical and hallucinatory cases can share the same internal modification while differ in external cause and success.

The view is ontologically economical and preserves a common factor. It also avoids the homunculus temptation: the subject does not inspect an internal display. The experiential event is the subject's state.

Its central problem is giving the state enough structure. Natural language adverbs hide relations among qualities, locations, and objects that perception clearly distinguishes.

The many-property problem

Suppose a subject experiences a red square beside a blue circle. A simple adverbial analysis says the subject senses redly, squarely, bluely, and circularly. That description does not distinguish the intended arrangement from a blue square beside a red circle. The properties need to be bound correctly.

One can introduce complex adverbs such as red-squarely-and-blue-circlely. But an independent primitive may be required for every possible configuration, losing compositional explanation. Alternatively, adverbialists can add structured manners of sensing, with positions and quality bindings. The more structure is added, the more the account resembles representational content.

The many-property problem shows why avoiding objects is not enough. Experience presents an organised field, and a theory must explain its compositional structure.

Particularity

Perception appears to concern particular objects, not only arrays of qualities. A subject tracks this cup through changing angles and distinguishes two qualitatively identical cups. Sense-data can be particulars but make the tracked external object indirect. Simple adverbialism supplies qualities without an obvious account of particularity.

Demonstrative or object-file representations offer a contemporary solution: perceptual systems maintain representations tied to individual objects across change. This is an intentional rather than sense-datum account. An adverbialist can incorporate object- directed structure, but again moves toward structured representation.

Causal intermediaries are not perceived intermediaries

Retinas, neural signals, and cortical models causally mediate perception. It does not follow that subjects perceive them. A camera sensor mediates a photograph of a tree without becoming the tree depicted. The existence of processing therefore supplies no direct argument for sense-data.

This point also protects direct realism from a crude objection. "Direct" perception need not be causally immediate. It means that ordinary objects, rather than private mental objects, are what perception presents or relates the subject to.

Sense-data require phenomenological and epistemic motivation beyond the existence of a causal chain. Once intentional content explains absent and misrepresented objects, that motivation is substantially reduced.

Contemporary descendants

Few contemporary theories retain classical sense-data as private objects. Their functions survive in other forms:

• internal representational states provide a common factor;
• phenomenal properties provide the experienced qualities;
• sensory formats provide spatial and qualitative structure;
• Bayesian or predictive models explain why perception can diverge from stimuli.

Adverbialism likewise survives as the insight that experience need not relate a subject to a mental object. Event-based and representational theories can treat qualities as features of experiencing while adding the compositional structure simple adverbs lack.

Assessment

Sense-data and adverbialism expose a genuine choice: explain appearance through an object of awareness or through a way the subject is modified. Contemporary views mostly reject private objects while retaining the demand for a structured common factor - or, as disjunctivism does, reject the common factor itself.

Selected references

• Chisholm, Roderick M. Perceiving (1957).
• Jackson, Frank. Perception (1977).
• Price, H. H. Perception (1932).
• Sellars, Wilfrid. "Empiricism and the Philosophy of Mind" (1956).