New Physical Properties
I. Properties, Physical Properties, and New Physical Properties
My aim here is to explore the hypothesis that there can be in the world, and that in fact there are, new physical properties. But, before doing it, we need to make clear the sense in which we are going to speak of properties, of physical properties and of the existence of new physical properties.
A. Referential Expressions for Properties
Strictly speaking, nominalizations and quotations are the usual procedures by means of which we get to construe referential expressions from the predicates of our languages. And it is through certain nominalizations, but not through quotations, that we get to construe from predicates something able to refer to properties. Which nominalizations would lead from predicates to referential expressions for properties? Consider a sentence like the following:
1. The sky has a salmon color.
There are three possible kinds of nominalizations that we could construe involving all the elements presents in 1:
2. The sky that has a salmon color.
3. The having a salmon color the sky.
4. The salmon color of the sky.
Here, the sky that has a salmon color is an object. And 2 is a referential expression for an object. Just for the same object which is referred by the subject of sentence 1. The having a salmon color the sky is an event. And 3 is a referential expression for an event. Finally, the salmon color of the sky is a property. And 4 is a referential expression for a property.
Let us focus on expressions like 4. We can call them subject-involving referential expressions for properties. They involve what is referred by the grammatical subjects of their associated sentences, 1 in our case. In other words, 4 involves just the object which would be referred by the referential expression 2. The sky involved in 4 is just the sky that has a salmon color. If we omit that component, we obtain from 4 the following referential expression:
5. The salmon color.
Nominal expressions like 5 also would be referential expressions for properties. In general, we can call them non-subject-involving ones. Every declarative sentence of a language is associated with one and only one subject-involving referential expression like 4. And every declarative sentence of a language is associated with one and only one corresponding non-subject-involving referential expression like 5.
According to that analysis, we could understand properties as the intended referents of the resulting expressions of nominalizing the declarative sentences of a language in the way 4 and 5 do. This is how we will understand that properties are intended to be identified and referred with the expressive means of our languages. First we need to have certain declarative sentences, then we nominalize and construe certain expressions able to refer to some properties. The more declarative sentences a language is able to offer, the more properties could be identified and referred through that language.
B. The Existence of Properties
In principle, all the expressions construed in the manner of 4 and 5 do intend to refer to properties. However, many intended referential expressions for properties are vacuous, where being vacuous entails that there does not exist any property at all that can be referred by them. How to tell more precisely the genuine referential expressions for properties from those that are not? There is one very simple and intuitive idea about what is the existence of properties that might be of help here. In short, that idea is that to say that a property exists is the same as to say that the property really can be instantiated. The existence of properties is the existence of possible instances. Even though no entity actually have the property in question, the property exists if it can be had by some entity. More, a property can exist even if in fact it never gets to be instantiated. With respect to properties, to exist is not to be instantiated but to be instantiable. We can use this idea to establish the following criterion:
A referential expression for properties, like 4 or 5, would be genuinely referential if it is possible to follow the opposite direction to our previous nominalizations and to reconstrue from it, or from it and some other referential expression taken as subject, certain well formed declarative sentence which is not necessarily false.
Now we have become to a more refined position about properties. Properties identifiable through a language would be the referents of the resulting expressions of nominalizing the declarative sentences of that language in the way of 4 or 5. The availability of such nominalizations would be necessary in order to identify properties and refer to them. However, it would not be sufficient. These nominalizations must be made in a way that from them, or from them and some other referential expressions taken as subject, some well formed declarative sentences not necessarily false can be reconstrued. In our scientific and ordinary languages, the existence of properties is linked to the possible truth of certain declarative sentences and the nonexistence of properties is linked to the bad formation or necessary falsity of other declarative sentences.
C. Co-referentiality and Identity
To discover the co-referentiality of some of the referential expressions for properties we are using is to discover the identity of the properties referred. And there is one special case of co-referentiality that will be worthy of mention here. It occurs when there is some kind of, let us say, compensation between the specifications present in some non-subject-involving referential expressions and the specifications present in other subject-involving ones. Consider, for instance, the following two referential expressions for properties:
6. The having a shining salmon color with purple shades and delicate metallic flashes.
7. The having an extraordinary color of the sky of Mars summers sunsets.
The specifications present in 6 are very rich. But 6 lacks subject-involving elements. In contrast, the subject-involving elements of 7 are very rich, whereas the other specifications are much more generic. In these situations, sometimes we can have referential expressions which are co-referential. The lack of subject-involving elements in one referential expression can be compensated with a richer specification in the other elements, and vice versa.
This is just what we can find in our examples. We can have reasons to consider that 6 is co-referential with 7. They would have to be considered as co-referential in the case we would have good reasons to think that the extraordinary color had by the sky of Mars summers sunsets could not be other but the shining salmon color with purple shades and delicate metallic flashes. If that were the case, then the following identity statement would be true:
8. The having a shining salmon color with purple shades and delicate metallic flashes is the having an extraordinary color of the sky of Mars summers sunsets.
Now, compare 8 with the following two identity sentences:
9. The blue is the color of the sky.
10. The being water is the being H2O.
The identity expressed in 9 seems to be, if true, a very contingent identity, a simple question of fact. In contrast, the identity expressed in 10 seems to be a necessary one, a theoretically grounded identity. Sentences like 8 would be placed between these two poles. Sometimes they would express mere contingent identities. But other times they could express identities very close to the identities expressed by something like 10. Without any direct theoretical grounding, the compensation of specifications in 8 really can produce a modal force as strong as the one had by 10.
That semantical mechanism of compensation offers an important and very often unnoticed way to refer to properties with respect to which we do not have in our languages the expressive means to construe powerful non-subject-involving specifications like 6. We could easily refer to them through other, let us say, subject-involving referential expressions like 7. And this could be the case with respect to some physical properties. Even if physics were not able to employ powerful non-subject-involving referential expressions to refer to these physical properties, we could refer to them through certain other subject-involving referential expressions beyond the language of physics.
D. Kinds of Referential Expressions and Kinds of Properties
Even though several referential expressions can refer to the same property, they do not have to be synonymous. And they do not have to be involved in the same epistemic and linguistic contexts either. Think, for instance, in the following two referential expressions:
11. The having certain acceleration of this car.
12. The having certain acceleration of exactly 7,0124 meters per second.
Suppose that both expressions really are co-referential and that they display the adequate compensation between specifications as to suggest that we are dealing with a necessary identity. This would amount to the necessary truth of the identity sentence:
13. The having certain acceleration of this car is the having certain acceleration of exactly 7,0124 meters per second.
By assumption, 11 and 12 would be co-referential. However, only 12 would be of help in order to make the appropriate inferences to predict, through some physical laws, that the car in question will be able to move 70,124 meters in 10 seconds. Even though 11 and 12 are co-referential, the referential expression 12 has a place in our physical explanations and predictions that cannot be occupied by 11.
Here, the important point is that, assuming the co-referentiality of 11 and 12, if the property referred by 12 is a physical property, then the property referred by 11 has to be also a physical one. Both expressions do refer to one and the same property. In spite of that, only 12 is a referential expression that can belong to the languages of physics. So, with respect to the same property, we can have referential expressions which belong to certain languages or theories and referential expressions which do not belong to these languages or theories. Moreover, we can have referential expressions only of the second kind.
E. Physical Properties
When do we consider that a property is physical? One plausible answer would be that a property is physical in the case its instances have effects of a kind of effects which we acknowledge as physical in a non trivial sense. The following criterion tries to do justice to that idea:
A property P has to count as physical when we know there is at least one relevant physical kind of events K such that at least one of the effects that can be caused by each instance of the property P belongs to kind K.
A physical kind of events would be a kind of events such that 1) they instantiate the same physical property, or 2) they instantiate different physical properties but they entail the existence of other events in which the same objects than before instantiate the same physical property at the same time than before. We need a further qualification. The physical kinds of effects mentioned in our criterion would have to be relevant ones. The question is that whatever physical effects the instantiation of a property can have, these effects would entail events in which the same physical property is instantiated. By this way, every property having any kind of physical effects would have to count as a physical property. But we cannot be so tolerant. The relevant physical kinds of effects would have to be those which show a projectable character over particular physical causal stories.
In order to consider a property as physical, it would be sufficient to know that at least one of the effects that can be caused by each instance of the property belongs to a relevant physical kind of events. At least one of such effects has to permit the identification of a particular and projectable physical causal story. The instantiations of physical properties have to have effects integrated in the net of particular and projectable causal stories we assume as being part of the whole causal history of the physical world. They have to make causal differences of at least one relevant physical kind. We know which referential expressions for properties do refer to physical properties through the effects it can have to instantiate those properties.
F. New Physical Properties
According to all of that, we can now define the notion of new physical properties as follows:
A property would be a new physical property from certain time (or portion of time) t iff
1) up to t, the property in question did not exist or, if it did exist, its possible instances did not have the appropriate causal effects, and
2) from t the property gets to have possible instances showing the appropriate causal effects.
Several important points would be worthy of consideration here. The first one would be that not every new physical property have to be a new property. The second point would be that old physical properties could leave to be physical properties in just the opposite way to the way in which there can be new physical properties. The third point would be related with our referential means to identify properties. With respect to the new physical properties it would not be necessary to dispose of referential expressions which belong to the languages of physics, and it would not be necessary either to have referential expressions which are co-referential with referential expressions belonging to the languages of physics. The fourth and last point, but not the less important, is that the physical effects which serve to qualify some properties as physical can be themselves new physical events which did not exist before. Moreover, with respect to these physical effects, we can be in the same situation than with respect to the new physical properties in general.
II. Four General Situations of Physical Novelty
Now, we are going to describe four general situations in which there would be a strong pressure to admit the existence of new physical properties. We will make use of the criterion previously stated that if the having certain property shows effects belonging to certain relevant physical kind, then the property has to be considered a physical one. Also, we will use the procedure previously described of referring to a property through subject-involving referential expressions.
A. First Situation: Causal Structures
This situation would be based on the existence of causal relationships which are themselves the effect of other structuring causes. Let us suppose the following general scenario:
(1) Some causal relationships between events exist if and only if they are the effect of certain causes which have the structuring power to get that the instantiation of some properties have a causal control over certain relevant physical kinds of effects.
(2) Some events have caused in time t the causal relationships C1.
(3) The having the property F can cause the relevant physical kind of effects of having G if and only if there exist C1.
Here, the property F would have to count as a physical property from time t. The causal relationship established in t between the having the property F and the having the relevant kind of physical effects of having the property G makes that from t the property F is a physical one. Only after t, but not before, the property F can have instances with effects belonging to a relevant physical kind.
B. Second Situation: Thresholds
This situation is based on the fact that certain relevant physical kinds of effects can depend on the existence of certain other effects related with the former ones in an accumulative way. These other effects would be like a threshold with respect to the former relevant kinds of effects. Let us suppose the following scenario:
(1) There are some relevant physical kinds of effects that exist if and only if other effects have taken place in an accumulative way.
(2) Certain events have caused in time t the accumulative effects C2.
(3) The having F causes the relevant physical kind of effects of having G if and only if the accumulative effects C2 have taken place.
Again, as in the first situation, only from time t the property F would have to count as a physical property.
C. Third Situation: Triggers
This situation is symmetrical with respect to the second one. It is like to have a threshold by the side of the causes. It is based on the fact that the existence of certain relevant physical kinds of effects of the instantiation of some properties can depend on the accumulative instantiation of other properties related with the second ones. The instantiation of these other properties would be like a trigger with respect to the former relevant kinds of effects. Now, the scenario would be as follows:
(1) There are some relevant physical kinds of effects that exist if and only if their causes have been instantiated after the instantiation of other properties in an accumulative way.
(2) In time t it has taken place the accumulative instantiation of properties C3.
(3) The having F causes the relevant physical kind of effects of having G if and only if the accumulative instantiation of properties C3 has taken place.
Again, as in the first and second situations, from time t the property F would have to count as a physical property.
D. Fourth Situation: The Background
The four situation would be based on the fact that certain relevant kinds of effects can depend on certain peculiar conditions in the immediate background. To construe the scenario, let us suppose the following:
(1) There are some relevant physical kinds of effects that exist if and only if certain background conditions take place.
(2) Certain events have caused in time t the background conditions C4.
(3) The having the property F cause the relevant physical kind of effects of having G if and only if the background conditions C4 are present.
In this situation, again, the property F would have to count as a physical property from time t. Before conditions C4 have taken place in time t, property F was not a physical property. But once the background conditions C4 do occur, the property F is able to have the relevant physical kind of effects of something being G. And this makes F to become a physical property.
We have presented four situations in which 1) our criterion for a property to be physical, together with 2) the use of subject-involving referential expressions for properties, and with 3) the existence of certain peculiar kinds of physical effects forces us to accept the existence of new physical properties. Is our typification exhaustive? I have no clear idea about that question. It is true that, in certain cases, some of our four situations could be considered as reducible to some of the other ones. The background is particularly able to subsume any of the other situations. Anyway, all of them would deserve consideration by their own. And general appeals to the background would be in many cases very few informatives.
Another point seems to me here more interesting. The four situations could overlap at many points. Causal structures, thresholds, triggers, and background conditions can be combined in many ways. That is what would happen, for instance, in we combine the situations and examples above presented as follows:
Structural features and events in this sort of device cause that when that lever is moved from A to some other position, or from some other position to C, then its motor improves its working power. But that movement of the lever also makes that certain metallic piece becomes to be placed more and more close to the proximities of the tank of water. So, when the motor is hot, and the environmental conditions are favorable, the temperature of the water can increase very fast reaching suddenly the threshold of 100º C.
Our books of engineering and applied sciences are full of descriptions like that. They appear also in psychology, social sciences and everyday discourse. Complex behaviors, feedback relationships, control mechanisms, delays effects, etc., are made from situations like the four ones we have presented above. The descriptions of these phenomena are very peculiar in that they are always associated with subject-involving referential expressions for properties. And we cannot forget that peculiarity. If we think that those subject-involving referential expressions are simply co-referential with the corresponding non-subject-involving ones, then we loose the particular and projectable causal stories we are dealing with. And without these stories, we could not appreciate that some relevant kinds of physical events can force us to consider that we are referring to some new physical properties which perhaps cannot be identifiable and referred other ways.
Why it has been supposed so often that there do not exist new properties in the physical world? Almost all the philosophical positions which accept the ontological primacy of the physical assume that there are not new physical properties, or they simply ignore that possibility. Even accepting that for a property to exist it has to be instantiable, and even accepting that for a property to be physical its instantiations have to show effects of some relevant physical kinds, they would reject the appearance of new physical properties in the four situations previously presented. Here, I think that we are faced with two mistaken conceptions, the one having to do with the ways we can refer to properties and the other having to do with the notion of being instantiable.
We have insisted in the importance of our subject-involving referential expressions for properties. And we will not repeat here our arguments. Subject-involving referential expressions for properties would offer the expressive means to refer to properties that perhaps might be not identifiable through any other mean. The strong tendency to assume that the predicates of our languages directly refer to properties is the responsible of not having taken into account the decisive referential role played by that sort of referential expressions. The other mistaken conception, which concerns to the notion of being instantiable, would consist in supposing the following:
If it is true from certain time t that a property can have instances with effects clearly belonging to a relevant physical kind, and so being a new physical property from t, then it has to be also true before t that the property could have had instances with effects clearly belonging to that relevant physical kind, and so being an "old" physical property before t.
But this is clearly wrong. And the reason is that the possibility to have instances having causal effects belonging to a relevant physical kind becomes less and less possible conforming that possibility depends on other possibilities which can or cannot be actualized. It is in that sense in which our previous criterion for a property to be physical has to be understood. The effects that can be caused by the instances of a property become less and less possible effects of such a property in the case those effects also depend on the actualization of other conditions and properties which can or cannot be in fact instantiated. In logical terms, the possibility to have instances showing causal effects belonging to a relevant physical kind would have a non-transitive modal behavior. It would behave like the operator of possibility in modal system T. Or, in any case, it would behave like the operator of possibility in some system between T and S4. The last would be necessary if we want to say that the properties F, in the four situations, would exist not exactly from the time (or portion of time) t in which the structuring events, the thresholds, the triggers, or the background conditions do exist, but from the time they are "close" to exist. (By the way, to appeal to the total cause of certain event instead of to the partial causes of it would not change anything of what we are saying.)
III. Some Consequences of Physical Novelty
The hypothesis of the existence of new physical properties would have very interesting consequences with respect to phyisicalism, reduction, special sciences, the layered image of reality, multiple realizability, emergence, downward causation, and so on. (About these topics, see specially the references at the end of the paper)
That hypothesis would make possible to distinguish a foundationalist physicalism from a non-foundationalist one, and to articulate the idea of the ontological primacy of the physical in a way that it would not depend on an ultimate level of physical properties given once and for all. In this sense, the possibility of new physical properties also would have immediate consequences for reduction. New physical properties would not be reducible, they would not be eliminable and to accept them would not have to motivate necessarily any dualism.
The proper understanding of the special sciences, and the layered picture of reality usually associated to them, also would be affected by the acknowledge of the existence of new physical properties. In particular, it would offer another ways to look to the problems originated by the phenomenon of multiple realizability and by the notion of emergent properties and downward causation. In certain circumstances, multiple realizable properties simply could become new physical properties. Properties belonging to some level could go downstairs to other levels below.
Also, the prospect for the so-called emergent properties would not only be conservative reduction, elimination or some kind of dualism. These options would not be all the available options. There would be another unexplored and promising possibility. In all our paper, we have take care not to use the notion of emergence. But we have tried to show how there could be new properties in the physical world. This way, some of the so called emergent properties could simply become new physical properties. And, therefore, some downward causations from the emergent to events placed in other levels below could become same-level causations. The physical basis of the world would not be something fixed once and for all. The possibility of new physical properties would offer a physicalistic picture of the world much less foundationalistic and much more dynamical than the one offered by classical physicalism. All the above ideas would have very special implications in some recent problems discussed in philosophy of mind.
Many of the new physical properties referred by these subject-involving referential expressions would have a physical character very unstable. But other ones could be more permanent through wide periods of time and with respect to many kinds of systems. And, among these, the new physical properties referred with the help of subject-involving referential expressions full of mental predicates could occupy a very important place. In any case, the set of physical properties would not be established once and for all. This is how the physical world would be made. At least, the physical world we live in.
Beckermann, A., H. Flohr & J. Kim, 1992, Emergence or Reduction, Berlin, Walter de Gruyter.
Kim, J., 1992, "`Downward Causation' in Emergentism and Non-reductive Physicalism," in Beckermann et al. (1992).
________, 1993, Supervenience and Mind, Cambridge, Cambridge University Press.
Putnam, H., 1975, "On Properties," in Philosophical Papers, vol.1., Cambridge, Cambridge Univ. Press.
Sabates, M., 1997, Mental Causation: Property Paralelism as an Answer to the Problem of Exclusion, PhD. Dissert., Brown Univ.
Searle J., 1992, The Rediscovery of Mind, Cambridge, MIT Press.
Yablo, S., 1992, "Mental Causation," The Philosophical Review, vol. 101, n. 2.