1. Introduction
In Natural kinds and conceptual change, Joseph LaPorteFootnote 1 puts forward an innovative view of natural kind terms, according to which the use of such terms is vague. Following a well-known linguistic tradition, he calls the hidden vagueness in the use of a term, which is later exposed as our knowledge grows, ‘open texture’.Footnote 2 Natural kind terms are then subject to meaning-change over time, whereby meaning-change should not be understood as synonymous with conceptual-change of the Kuhnian type, but instead as precisification of kind terms with open texture. LaPorte combines precisification with a Kripkean defence of the rigidity of kind terms. Any kind term rigidly designates the same kind in all possible worlds, although the extension of the kind term in question may change from one world to another, and even within the same world before and after a scientific revolution. Since there is an element of semantic stipulation as to how we refine the use of a vague kind term, LaPorte takes kind terms as rigid de jure, but not de facto.Footnote 3
Under the Kripke–Putnam view, rigidity accomplishes the other important job of getting kind terms hooked up to their referents. For example, the term ‘water’ refers to whatever possesses some underlying properties and relations that – empirically discovered – guarantee sameness of kind to paradigm water. But LaPorte argues that this is instead the role of the causal theory of reference and not of rigidity, and accuses Putnam of conflating the two.Footnote 4
Against Putnam's causal theory of reference, LaPorte maintains that possessing underlying properties and relations does not guarantee sameness of kind to paradigm samples. For example, being H2O is neither necessary nor sufficient for being water; nor did we discover that water is H2O. On the contrary, if confronted with XYZ we might conclude that ‘XYZ is water’; or vice versa, if confronted with H2O having unusual characteristics, we might conclude that ‘Some H2O is not water’.Footnote 5 LaPorte concludes that the kind term ‘water’ is vague and changes its meaning over time, so that XYZ is neither clearly in nor clearly out of the extension of the term until a decision is made by speakers to include or exclude it. Such stipulation on behalf of speakers would amount to a precisification of the term; it would not be a scientific discovery. To illustrate this point, LaPorte revisits Putnam's Twin Earth story with the fictional scenario of Deuterium Earth.Footnote 6 The story is functional to LaPorte's point that scientists stipulated that deuterium oxide (D2O), i.e. one of the most relevant isotopic varieties of water, is a kind of water (usually called ‘heavy water’). But they could have chosen otherwise, and they would not have been wrong in doing so, anymore than we are right in including deuterium oxide in the extension of ‘water’. Let us call this the thesis of semantic stipulation.
LaPorte's argument draws on the mixed nature of the liquid that we normally call ‘water’. Naturally occurring water is not 100% H2O. For example, ocean water contains also a small percentage of deuterium oxide (of the order of 0.015%, upon LaPorte's indication), without counting percentages of other isotopic varieties (such as tritium oxide, oxygen-18, etc.), as well as salt and gold, among other substances. Following a strategy already deployed against Barbara Abbott'sFootnote 7 Putnamian defence of water being H2O, LaPorteFootnote 8 tacitly uses the mixed nature of any ocean water sample as an argument against the claim that H2O content is necessary and sufficient to determine the reference of the term ‘water’.Footnote 9
In this paper, I object to LaPorte's Deuterium Earth story on two main grounds. First, scientific history belies LaPorte's intuitions as to why the Earthlings would have good reasons for not counting the Deuterium Earth's ocean liquid as a kind of water. Second, if we take LaPorte's view seriously, an unwelcome Goodmanian scenario may arise. In Section 2, I briefly review LaPorte's Deuterium Earth story and conclude that for the thesis of semantic stipulation to go through, one would need to prove that ‘dwater’ (as the Earthlings dub the Deuterium Earth's liquid) is a vernacular kind term on equal (scientific and semantic) footing with ‘heavy water’. Or better, one would have to prove that the two options:
(A) D2O is not a kind of water (let us call it ‘dwater’, as the Earthlings call it)
(B) D2O is a kind of water (let us call it ‘heavy water’, as the Earth scientists call it)
are on a par, and in particular that option (A) is a live option with equal semantic status and scientific credibility as option (B). But I contend that it has neither. In Section 3, I show that option (A) does not have the same scientific credibility of (B), because D2O does bear after all the same microstructural-kind relation to what we ordinarily call ‘water’. In Section 4, I show that option (A) does not enjoy an equal semantic status with option (B) either, because for ‘dwater’ to be a genuine kind term, it would have to be projectible, i.e. it would have to support successful inductive inferences. But, as it happens, the term is non-projectible and vulnerable to an unwelcome Goodmanian scenario. In Section 4.2, I present a series of possible responses to the Goodmanian scenario and I conclude that none of them successfully evades it.
2. LaPorte on Deuterium Earth
LaPorte's thesis of semantic stipulation derives its compelling force from his apt variation of Putnam's Twin Earth story. LaPorte presents us with the fictional scenario of Deuterium Earth visited by the Earthlings in 1905.Footnote 10 The space travellers land on Deuterium Earth equipped with samples of various biological and chemical substances from planet Earth, which they soon begin to compare with the substances populating Deuterium Earth. The water-like liquid that fills the ocean, and that looks prima facie identical to the Earthlings' sample of water from planet Earth, reveals soon some dissimilarities even before molecular testing is done: no fish or aquatic plant can live in it and its boiling and melting points differ from the ones of water.
The Earthlings dub the non-better-identified microstructure PQR and find also traces of PQR in their own ocean-water sample from planet Earth (in the order of 0.015%). ‘Eventually, through sophisticated testing’,Footnote 11 the Earthlings find out that the water-like liquid consists of hydrogen atoms, whose mass is greater than the mass of ordinary hydrogen atoms, because of one neutron (in addition to the proton) in the nucleus. In honour of Deuterium Earth, they call the new element ‘deuterium’.
The Earthlings learn from the Deuterium Earthlings that an explosive weapon was built with deuterium and that the explosion caused a whole island to vanish.Footnote 12 No such bomb is known to the Earthlings. Thus, in the light of all these differences, the Earthlings decide to dub the water-like liquid that fills the ocean on Deuterium Earth ‘dwater’, so as to distinguish it from ‘water’.
Thirty years elapse: it is the year 1935 when the Earthlings at last return to planet Earth, bringing with them samples of substances from Deuterium Earth. When confronted with the ocean-sample of ‘dwater’, the scientists on Earth respond disappointingly that it is not a new liquid but just a chemical variety of water, made up by the hydrogen isotope, which they had coincidentally called ‘deuterium’ to distinguish it from protium (i.e. the hydrogen atom, whose nucleus consists of one proton only). LaPorte draws the following moral from the story:
What do I think this story shows? I think it shows that we did not discover that deuterium oxide is water. Hence we did not discover that water is identical to H2O. We could have concluded that some H2O (the variety that is D2O) is not what we had been calling ‘water’, as our space travellers concluded. I think the decision that just H2O made with protium bears the same microstructural kind to the majority of what we called ‘water’ would have been no less acceptable a conclusion than that H2O bears the key relation. We cannot say that our space travellers were just flat wrong in concluding that D2O is not what they had been calling ‘water’ and that we are just plain right in concluding that it is.Footnote 13
In sum, ‘water’ is a kind term with open texture and it is a matter of semantic stipulation whether we decide to include or not deuterium oxide in its extension. We could have gone either way. Or, so LaPorte argues. This is what I call the thesis of semantic stipulation.
Alexander BirdFootnote 14 attacks LaPorte on this specific point by noting how chemistry is the science of substances, and that although the space travellers might have been right in not counting D2O as water, their reasons were practical rather than chemical: ‘if you want to know what a substance is rather than what it will do to your body or whether you can make a fusion bomb from it, you turn to a chemist’.Footnote 15 Moreover, Bird draws attention to the use of vernacular kind terms as opposed to natural kind terms and to Putnam's division of linguistic labour to conclude: ‘the claims about open texture are most plausible concerning vernacular terms that appear to be natural kind terms. But important identity statements in science are not just those that conjoin a vernacular kind term with a scientific one’.Footnote 16
I follow up on Bird in arguing against LaPorte that it is not the case that we could have gone either for the Earthlings' option of not including D2O in the extension of water, or for the Earth scientists' reverse option. In other words, it is not the case that the two options are on a par, i.e. that whether or not D2O is a kind of water is ultimately a matter of semantic stipulation. My line of argument departs from Bird in two respects: (1) I think that LaPorte is after all correct in claiming that the Earthlings' reasons for taking D2O not as a kind of water are ultimately of chemical (as opposed to practical) nature; and (2) I also think that LaPorte is correct in stressing the importance of vernacular kind terms with open texture for scientific claims (in particular, I am thinking of the role of vernacular kind terms for inductive inferences, as opposed to theoretical identity statements).
Whilst I concede these two points, I do not think they jointly license the thesis of semantic stipulation. For the thesis to be defensible – out of the seemingly compelling force of the fictional scenario – we need to assess whether the Earthlings' conclusion that D2O is not a kind of water but instead a new chemical substance (called ‘dwater’) stands up scientific and semantic scrutiny. In the next two Sections, I cast doubt on both.
The two points are related. For ‘dwater’ to have the semantic status of a genuine kind term, it would have to be projectible: i.e., it would have to support successful inductive inferences. But I am going to argue that the term is either non-projectible; or its projectibility depends ultimately on facts about the D2O content of the new liquid as an isotopic kind of water. I must then first show that facts about the D2O content of Deuterium Earth's ocean liquid underwrite the conclusion that it is an isotopic kind of water.
3. Why ‘dwater’ does not have the same scientific credibility as ‘heavy water’. The Deuterium Earth story continued
In this Section, I show that the Earthlings are in fact wrong in thinking that D2O is not a kind of water. Facts about the D2O content of Deuterium Earth's ocean liquid underwrite its nature as an isotopic kind of water; and there are reasons – I think – for resisting LaPorte's conclusion that some H2O (for example, the D2O variety of it) is not what we have been calling ‘water’. Let me continue the Deuterium Earth's story where LaPorte's one ends, following this time the historical records about the discovery of deuterium, and the ensuing scientific debate.
It is the year 1935. The Earthlings have at last returned to planet Earth after more than thirty years field-work on Deuterium Earth. They present a jar of Deuterium Earth's ocean liquid to the Earth scientists, who upon investigation conclude that it is not a new liquid but just a new kind of water that they have themselves recently discovered and independently called ‘deuterium oxide’ or ‘heavy water’ (the first is the scientific kind term, the second is the vernacular kind term).Footnote 17 Lots of things have happened in the scientific world since the Earthlings left planet Earth in 1905 – the Earth scientists explain to the Earthlings.
‘To start with, in 1919 the isotopic constitution of elements became well-known. The term “isotope” was originally used by Frederick Soddy to designate chemically non-separable varieties of the same element, whereby in the case of heavy radioactive atoms the differences in atomic weight did not give rise to observable chemical differences. But the use of the term has changed since (despite Soddy's acrimonious defence of the original meaning): now (i.e. in 1935) the term “isotope” is ordinarily used to indicate atoms with the same nuclear charge but different atomic weight’.
‘This leads us’ – as the Earth scientists hasten to explain to the Earthlings – ‘to the rather delicate issue of finding a suitable name for the new liquid, following up on the discovery in 1931 of the hydrogen isotope containing a neutron in addition to the proton, by three American scientists.Footnote 18 Urey discovered the new isotope by evaporating liquid hydrogen at low temperatures, and examining the resultant spectral lines. His discovery followed of one year the discovery of two other isotopes of oxygen (i.e., 17O, 18O in addition to the known 16O), on the basis of a measured discrepancy for the hydrogen atomic weight'.
The Earth scientists continue their report to the Earthlings: ‘The choice of a name has proved rather difficult and it has been the subject of lively discussion. The Times of 9 December 1933 announced in a column the discovery of “heavy hydrogen” as a “a new kind of water”, which has escaped discovery hitherto only because of its very small quantity in nature compared to ordinary hydrogen, and whose presence in ordinary water would significantly alter the ability of water to support life. The search for a suitable name was open and an entire meeting of the Royal Society was dedicated to it on 14 December 1933’.Footnote 19
‘At the meeting, Soddy, the father of isotopy, refused to consider this element as an “isotope” on the ground that it defied his original meaning of “isotope”, reserved – as it was – to chemically non-separable varieties of the same element.Footnote 20 But other physicists were of a different advice. Lord Rutherford, for example, retorted to Soddy that the term “isotope” was now extended to include chemically separable elements having the same nuclear charge and different mass, and he himself proposed the name “diplogen” for the new element. “Diplogen”, meaning in Greek double, found some enthusiastic reception among the scientists present at the meeting, because, in the words of Sidgwick, “the interesting thing about it is not that it is the second lightest particle….[but rather] that it is double the first one”.Footnote 21 Moreover, since the atomic volume and atomic number are the same as the hydrogen, the physical properties of this substance would not differ much from those of hydrogen, to the point that – as Dr Aston reported at the meeting – Prof. Bohr preferred to call it simply “hydrogen” since its atomic number was 1 and it was not a new element. Yet, there are some noticeable differences in the chemical behaviour of the new element, especially its rate of diffusion and chemical reactions, which were duly reported at the Royal Society meeting’.
‘Although these isotopic species of water molecules are physically and chemically very similar’, the Earth scientists continue, ‘they give rise to divergences at the macroscopic level of thermodynamic properties such as temperature of maximum density, heat capacity, volume and compressibility. For example, the temperature of maximum density, that is the temperature at which (∂ρ/∂T)p = 0 is 3.98 °C at atmospheric pressure for H2O, but becomes higher (of the order of 11.44 °C) for D2O. Hence, the different melting and freezing points of the two liquids. Those macroscopic changes are due to the complex dynamics of liquid water, where both intra-molecular vibrational modes and bending and stretching of the hydrogen-bonds and O–D bonds enter. And in biochemical reactions, replacing ordinary hydrogen with “heavy hydrogen” has far-reaching effects on the way living cells operate.’
‘Until a year ago (i.e. 1934), 2H still did not have a universally agreed name. Urey suggested “hydrogen two” for its nucleus and “pycnogen” for the element itself; Gilbert Lewis was in favour of taking 2H as an element in its own right called “dygen”, and called its nucleus “dyon”. Eventually, Urey opted for “deuterium”, upon suggestion of professors in the Dept. of Greek, and the name has now been universally accepted’.Footnote 22
‘It has also became clear’, the Earth scientists point out, ‘that given the very low percentage of “heavy water” contained in naturally occurring water, producing pure heavy water by electrolysis of ordinary water is a very difficult and expensive process. To produce one kilogram of heavy water, 50 tons of ordinary water have to be treated for one year, consuming 320.000 kilowatt hours, and, the output would still have a purity no better than about ten per cent’.Footnote 23 ‘The first industrial scale production of heavy water has just opened in Vemork, Norway, a year ago (in 1934)’, the Earth scientists continue, ‘and scientists are all excited at the prospects of its possible applications in biochemistry, and even its possible therapeutic uses. We are still working on possible methods of estimating the quantity of deuterium in naturally occurring hydrogen, following up on the works of Drs. L and A. Farkas, but all the evidence points to a very low percentage which could potentially create problems when it comes to industrial scale production.’
Here the report of the Earth scientists to the Earthlings on heavy water ends, as of 1935. The following events are now history. The nuclear plant in Vemork was bound to play a key role in the nuclear race that eventually led to the production of the H-bomb.Footnote 24 During World War II and throughout the 1950s and 1970s, scientists explored a variety of methods for production of heavy water on industrial scale, the two main ones being distillation processes from water or hydrogen, and chemical exchange processes involving either water-hydrogen or water-hydrogen-sulfide. In both types of processes, the production of heavy water faced a serious challenge due to the very low deuterium-to-hydrogen ratio in nature.Footnote 25
Ordinary water has indeed turned out to be a mixture of various isotopic varieties, such as 1H216O, 1H218O, 1H217O, 1H2H16O. Isotopic variation is not normally regarded as relevant for the purpose of calculating physical properties of water because those properties do not vary much on a molar basis.Footnote 26 But it does become relevant when dealing with macroscopic, thermodynamic properties dependent on the atomic weight of the molecules.Footnote 27
What lessons should we draw from the intricate (historical and scientific) details of the ‘Deuterium Earth story continued’? The first obvious lesson is the sheer historical contingency of dubbing and re-dubbing, echoing Hacking's.Footnote 28 That Urey's ‘deuterium’ was ultimately chosen over Rutherford's ‘diplogen’ or Lewis's ‘dygen’ is a sheer historical contingency, based on the tacit principle that the discoverer gets priority in naming. We might well have chosen ‘dwater’ to designate D2O. But one thing is certain. From these series of historical contingencies of dubbing and re-dubbing, it does not follow that we did not discover that D2O is a kind of water. Nor does it follow – as LaPorte recommends – that D2O content is neither necessary nor sufficient to fix the reference of both ‘dwater’ and ‘heavy water’. I take it that for a paradigm sample of D2O not to be baptised as a kind of water, but as some new liquid, a main condition would need be satisfied:
(I) the paradigm sample of D2O would have to bear nosame microstructural-kind relation to any ordinary sample of water in our planet.
But in fact, any paradigm sample of D2O would fail to meet condition (I.) on two distinct grounds:
i. isotopes, although not chemically identical, share nonetheless many physical properties, given the same atomic number – which is the reason why they do not occupy separate places in the periodic table (Bohr lost the battle of dubbing, but won the war on the periodic table).Footnote 29 Hence, given the same atomic number, and given our system of periodic classification, scientists are justified in counting isotopic varieties of water (such as D2O) as kinds of ‘water’.Footnote 30 The physical properties of D2O and H2O on a molar basis are sufficiently similar, and differences emerge at the macroscopic level of their thermodynamic properties (such as temperature of maximum density). On the basis of physical similarity at the molar basis, there is again no overwhelming reason for counting deuterium oxide as a distinct substance from ordinary water, which recall is not pure protium oxide but a mixture of several isotopic varieties (although LaPorte is correct in noting that at the macroscopic level of properties, including its melting and freezing points, it looks like a distinct substance). Then, D2O differs from ordinary H2O at the level of chemical macroscopic properties, but not at the level of the physical microscopic properties. But to count as a distinct substance not bearing same microstructural-kind relation, D2O would need to differ from ordinary H2O in terms of physical and chemical properties, which is not the case.
ii. Most importantly, naturally occurring water is a mixture of several isotopic species of both hydrogen and oxygen, whose very small percentages are not sufficient to alter the overall chemical and thermodynamic properties of ordinary water at the macroscopic level. This is the reason why isotopic varieties are normally included in the extension of ‘water’, and not classified as separate substances.Footnote 31 In particular, the very low D/H ratio in naturally occurring water means we will never encounter a Deuterium Earth's ocean, or any significant amount of naturally occurring deuterium oxide that could serve as a paradigm sample for the Earthlings' imagined baptism (because of chemical macroscopic differences with our ordinary H2O), pace LaPorte's fictional scenario. The only significant amount of D2O that we could encounter and that could function as a paradigm sample for the Earthlings' imagined baptism would be industrially produced via causal mechanisms involving either distillation or chemical exchange from ordinary water. But this would undercut the seemingly probative force of LaPorte's thought-experiment as presumably the quantities so produced would not count as naturally occurring in our actual world (which is a key aspect of any Putnamian-like baptism – see more on this point in Section 4.1).
Thus, either the Deuterium Earth's ocean sample does not meet condition (I.) (because of i.); or it does meet condition (I.) at the cost of not being a paradigm sample of deuterium oxide (because of ii.). Ordinary D2O does bear same microstructural-kind relation to ordinary H2O to deserve The Times's appellative of ‘a new kind of water’.
LaPorte suggests that the Earthlings' conclusion that D2O is not a kind of water is justified by the absence of same microstructural kind relation to what we ordinarily call ‘water’,Footnote 32 given that the liquid does not support aquatic forms of life, has a different melting and freezing point, etc. But as it happens, at the bare microstructural level and physical properties at the molar basis, there are no overwhelming differences between the physical properties of ordinary H2O and those of ordinary D2O. The differences arise at the macrostructural level, where thermodynamic properties enter, due to the complex dynamics of liquid water. Thus, to claim that D2O may not be counted as a kind of water on the basis of these macroscopic thermodynamic properties would not per se undermine the Putnamian claim that D2O does bear the same microstructural kind relation to the majority of what we have been calling ‘water’ all along.
But there is more. Most of what we have been calling ‘water’ consists of an isotopic mixture at the microstructural level, with tap waters from various countries easily differing by as much as 15 percent in the D/H ratio.Footnote 33 And in the absence of a single sample of average ocean water easily available to circulate widely, a new ‘standard mean ocean water’ (SMOW) was defined in relation to the National Bureau of Standards isotopic reference sample No. 1 (NBS-1), such that
$$\hbox{D/H \lpar SMOW\rpar }\equiv \hbox{1.050 D/H \lpar NBS-1\rpar }$$‘thus tying the standard to a sample readily available for world-wide distribution’.Footnote 34 With the new SMOW in place, it has been possible to measure the deuterium enrichments per millage in various samples of ocean water, ranging from the −0.7 of the Atlantic, to the +0.9 of the Pacific, and +0.1 of the Indian ocean, leading to an average D/H ratio of 1/6328.Footnote 35
The standards used to measure the deuterium enrichments of ocean waters might be regarded as conventional stipulations. But they do not make the inclusion of deuterium enrichments into the kind ‘water’ itself a stipulation, nor do they make the Pacific ocean more ‘dwatery’ than the Atlantic! On the contrary, our resilient use of the term ‘water’ to designate average ocean water and the scientific quest for a consistent measurement standard for its isotopic varieties speaks against the seemingly probative force of LaPorte's thought-experiment. If we take the term ‘water’ to designate the life-supporting liquid we are accustomed to from lakes, oceans, and rain, then given the very small (average) D/H ratio in naturally occurring water, we are justified in taking isotopic varieties as one kind of chemical substance that we have been calling ‘water’ all along.
Moreover, if we consider that pure D2O does not occur naturally, then the term ‘water’ used to designate Earth's ocean liquid, and the term ‘dwater’ used to designate Deuterium Earth's ocean liquid, would presumably both refer to isotopic mixtures, distinguished only by different ratios of D/H content (low on planet Earth, and high on Deuterium Earth). But then the difference between the Earth's ocean liquid and the Deuterium Earth's one would be just a matter of degree. And if the Earth's ocean liquid deserves the name of ‘water’ despite being an isotopic mixture, one may wonder why the Deuterium Earth's ocean liquid should be denied such an appellative.
To sum up so far, if at the microstructural level ordinary D2O and ordinary H2O do not differ that much and if the macroscopic differences (arising at the level of thermodynamic properties) are presumably supervenient on the physical properties at the microstructural level,Footnote 36 the case can be made for why facts about the D2O content of Deuterium Earth's ocean liquid underwrite the conclusion that it is, after all, an isotopic kind of water. Since LaPorte's story relies heavily on those facts, the Deuterium Earth story does not license the conclusion that we could have gone either way in taking D2O as a kind of water or not.
But it is not just the scientific credibility of the vernacular kind term ‘dwater’ (i.e. meaning ‘not-a-kind-of-water’) that is at stake here. As I am going to argue in the next Section, also from a semantic point of view, ‘dwater’ is not on a par with ‘heavy water’. Recall that in LaPorte's story both Earthlings and Earth scientists have apparently access to the same scientific information about the microstructural nature of deuterium oxide; they just happen to use two different vernacular kind terms to refer to D2O. LaPorte's story does not tell us whether the Earthlings prior to 1935 know or use the scientific kind term ‘deuterium oxide’. All we know is that they call ‘dwater’ what the Earth scientists call ‘heavy water’. Let us grant that they both refer to the same microstructure, namely D2O, but they diverge about vernacular kind terms.
I contend that the two vernacular kind terms are not semantically on a par, because vernacular kind terms are opaque, so to speak, under coextensive predicate substitution.Footnote 37 Despite both ‘dwater’ and ‘heavy water’ being coextensive with the predicate ‘being D2O’, it is not the case that we can substitute ‘dwater’ for ‘heavy water’ in any sentence in which the latter appears (e.g. in inductive inferences) and preserve the projectibility of the term. The reason is that despite having the same extension (D2O), ‘dwater’ and ‘heavy water’ have different intensions. ‘Dwater’ is the vernacular kind term given to D2O in a fictional baptism where all the available evidence about macroscopic properties (melting and freezing points, poisonous nature for living creatures, and so forth) suggests that the substance in question is not a kind of water. Vice versa, ‘heavy water’ is the vernacular kind term given to D2O in the real baptism by The Times in December 1933, where the available evidence about the very low percentage of this isotopic variety in naturally occurring water suggested that the substance in question was a kind of water. Thus, co-extensivity notwithstanding, the different intensions translate into the two terms being opaque under co-extensive predicate substitution. Opacity has consequences for the projectibility of the terms. In particular, it gives rise to a dilemma. Either the two terms are not interchangeable because, as I am going to show in the next Section, while ‘heavy water’ is projectible, ‘dwater’ is not. Or, for them to be interchangeable in inductive inferences salva projectibility, ‘dwater’ would have to be construed as synonymous with ‘heavy water’, i.e. as a kind of water, undermining in this way LaPorte's story that ‘dwater’ denotes D2O as not-a-kind-of-water.
4. Why ‘dwater’ does not have the same semantic status of ‘heavy water’. A Goodmanian twist to the Deuterium Earth story.
Should the Earthlings then agree with the Earth scientists that the Deuterium Earth's ocean liquid is, after all, a kind of water? In the light of the discussion in the previous Section, and in the absence of two alternative systems of chemical classification for D2O in current chemistry,Footnote 38 it would prove hard to resist this conclusion. Let us grant then that the Earthlings eventually agree with the Earth scientists, upon hearing their report in 1935, that D2O is after all, a new kind of water. But they insist on preferring their own vernacular kind term ‘dwater’, as opposed to ‘heavy water’, to mark the significant differences between the two liquids at the macrostructural level. The imagined dialogue may go as follows:
Earth scientists: ‘So, you can see that the Deuterium Earth's ocean liquid is after all a kind of water in the light of its physical properties on a molar scale, and very low ratio in naturally occurring isotopic mixtures’.
Earthlings: ‘Fine. But we really feel that your vernacular kind term “heavy water” under-represents the very peculiar chemical features of the new liquid, and creates the potentially misleading association with what we ordinarily call “water”. After all, “water” is associated with a life-supporting liquid, whereas this liquid here is poisonous to all higher forms of life. We'd rather stick to our original term “dwater” to demarcate the chemical difference between the two. The two vernacular kind terms “dwater” and “heavy water” are on a par from a semantic point of view. They both refer to the same stuff, namely D2O. Yet they mark an important difference. “Dwater” refers to D2O not as a kind of water, i.e. it refers to D2O with respect to those aspects in which it differs from what we ordinarily call “water”.’
Earth scientists: ‘Do you mean with respect to the relevant thermodynamic differences at the macroscopic level, such as temperature of maximum density, and hence boiling and freezing points, and its implications for glucose oxidation and ATP synthesis?’
Earthlings: ‘Yes, precisely. We feel that those relevant differences deserve a brand new name, at a sufficient distance from “water” and associates (i.e. “heavy water”). Had we been at the Royal Society meeting in 1933 when the issue of the name was highly debated, we would have pushed for this alternative option. After all, it is only a semantic stipulation whether we include D2O into the existing kind term “water” (and just add an adjective to it) or we coin a new name for it. Semantically, “dwater” is on equal footing with “heavy water” and could be used interchangeably with it: it could feature into the same inductive inferences'.
Earth scientists: ‘Unfortunately, you are late in the dispute about naming. The choice has already been made.’
Earthlings: ‘But things could have gone otherwise. Had our expedition come back two years earlier, we would have lobbied for the alternative name “dwater” and won the battle at the Royal Society meeting, where, from what we have heard, some scientists like Aston shared our feelings.'
What should we make of the Earthlings' thesis of semantic stipulation? To be defensible, one must prove that the two vernacular kind terms ‘dwater’ (as the Earthlings dub the Deuterium Earth's liquid) and ‘heavy water’ (as the Earth scientists call it) are semantically on a par. But, in fact, if we take LaPorte seriously – i.e. that there is no Putnamian fact of the matter as to whether deuterium oxide is or is not a kind of water (i.e. that D2O content is neither necessary nor sufficient to fix the reference of the term, and that it is only a matter of refining our vague use of kind terms) – then an unwelcome Goodmanian scenario may arise.
ProjectibilityFootnote 39 is a distinctive feature of natural kind terms under a variety of alternative accounts (from Boyd's and Kornblith's realism about kinds, to Hacking's nominalism, and Kuhn's constructivism).Footnote 40 Kind terms must be projectible for them to be the basis of successful inductive inferences in science.Footnote 41 Thus, for ‘dwater’ to be semantically on a par with ‘heavy water’, one must show that it is projectible, i.e. that it supports successful inductive inferences. But ‘dwater’ is in fact vulnerable to the new riddle of induction, as the following Goodmanian twist to LaPorte's story shows. In Section 4.2, I consider some possible responses to it, and conclude that none of them successfully evades the Goodmanian scenario.
4.1. A Goodmanian twist to the Deuterium Earth story
Let us go back to the beginning of our story. It is the year of the expedition (1905) and the Earthlings have just landed on Deuterium Earth with a jar containing ocean-water. The liquid inside the jar looks prima facie identical to the liquid that fills the ocean on Deuterium Earth. But they soon realise that there are some important differences between the two. They decide to dub the non-better-identified microstructure of the Deuterium Earth's liquid as PQR. Eventually, through sophisticated testing, the Earthlings find out that the ocean-liquid consists of an oxygen atom and two hydrogen atoms that happen to have a bigger mass than the hydrogen atoms known to them, because of an additional neutron in the nucleus. They aptly dub the chemical element ‘deuterium’ in homage to Deuterium Earth, re-dub accordingly the microstructure of the ocean-liquid as D2O, and the ocean-liquid itself as ‘dwater’ to mark the difference from ‘water’ of planet Earth. So far, this bit of the story follows exactly LaPorte's. Here is now the Goodmanian twist. By dubbing the ocean-liquid ‘dwater’, upon inspecting several samples, the Earthlings draw the following inductive generalization:
(A) The liquid that fills the ocean on Deuterium Earth is dwater.
Many years elapse. It is the year 1935: the Earthlings return at last to planet Earth, with a jar containing the ocean-liquid from Deuterium Earth. The Earth scientists compare the liquid with samples of a new chemical element they have just identified as deuterium, and whose combination with oxygen gives deuterium oxide or ‘heavy water’, as they call the new isotopic kind of water. By comparing several samples, the Earth scientists conclude
(B) The liquid that fills the ocean on Deuterium Earth is heavy water.
The Earth scientists go on to explain to the Earthlings all the recent events concerning the discovery of the new isotope, as per Section 3 above. The Earthlings, faced with the new pieces of evidence eventually agree with the Earth scientists that from a scientific point of view, it is after all a new kind of water. But they insist on using their old terminology ‘dwater’ to mark semantically the difference from ordinary water, which they perceive as blurred by the term ‘heavy water’. They also claim, as we saw in Section 4, that ‘dwater’ is semantically on equal footing with ‘heavy water’: they can both be used interchangeably in inductive inferences. It is after all just a matter of semantic stipulation (as opposed to discovery) – they argue, along LaPorte's lines – that the new liquid ended up being classified as a kind of water.
Now, if this were the case, the Earthlings would be vulnerable to the new riddle of induction. The term ‘dwater’ would be like Goodman's grue: something is dwater if and only if it is first examined prior to time t (say, 1935) and found to be not-a-kind-of-water; or, it is examined after time t and found to be a-kind-of-water. If there is no fact of the matter as to whether D2O is or is not a kind of water and if D2O content is neither necessary nor sufficient to fix the reference of the term, and we could have gone either way – then the Earthlings (who in 1935 are likely to accept the Earth scientists' conclusion as per Section 3) would be vulnerable to drawing ‘gruified’ (or better ‘dwateried’) inductive inferences like (A), whereby at time t (namely, in 1935) the same evidence that supports the Earth scientists' statement (B), namely that the liquid that fills the ocean on Deuterium Earth is a kind of water (called ‘heavy water’), supports also the Earthlings' parallel statement (A) that the liquid that fills the ocean on Deuterium Earth is not a kind of water (called ‘dwater’).
Let me clarify this point. If one endorses the Putnamian view that the meaning of a natural kind term is not fixed by what Putnam calls the stereotype,Footnote 42 i.e. the various beliefs speakers may associate with the term (for example, in the case of ‘dwater’, beliefs concerning its poisonous nature, its ability to be used in a weapon and so on), but it is instead fixed by the extension of the term (e.g. D2O),Footnote 43 then in every possible world (including Deuterium Earth), in which a substance bears same microstructural-kind relation to what we ordinarily call ‘water’ on planet Earth (intended in the way specified in Section 3 above, i.e. in the Earth scientists' sense of ‘water’ as including all its isotopic varieties), that substance would count as a kind of ‘water’. Under the Putnamian account, D2O content is necessary and sufficient to fix the reference of the natural kind term. Moreover, the Kripkean rigidity of natural kind terms is understood as following from the fact that when we (in the actual world) give the ostensive definition ‘this liquid (say, deuterium oxide) is a kind of water’, we intend that for every possible world W and every possible liquid x in W, x is a kind of water if and only if x bears the same microstructural-kind relation to the entity referred to as ‘this’ in the actual world (i.e., in the Earth scientists' world).Footnote 44 For Putnam, natural kind terms are rigid designators to the extent that the Putnamian meaning of ‘meaning’ is fixed by ostensive definition of paradigm samples in our actual world.
By rejecting Putnam's meaning of ‘meaning’, and by re-coupling Kripkean rigidity to some form of semantic stipulation, LaPorte suggests that natural kind terms can still be rigid designators de jure (not de facto), but whether or not this liquid (e.g. deuterium oxide) counts as a kind of water is ultimately a matter of speakers' decision: we decided to count it as such, but we might as well have decided otherwise (had we had the first encounter with this liquid that the Earthlings, for example, had on Deuterium Earth). However, the reasons why the Earthlings – in this hypothetical first encounter scenario – did not count this liquid as a kind of water seem to have nothing to do with any fact of the matter about the isotopic nature of deuterium oxide per se, and all to do with the Earthlings' beliefs about the poisonous nature of the liquid, its ability to be used in a weapon, and so on (what Putnam would call the stereotype of a natural kind term). Had the Earthlings' had alternative beliefs about this liquid, the liquid might have counted as a kind of water, after all.
Moral: once we dispense with the Putnamian view that there is a fact of the matter as to whether this liquid (deuterium oxide) here in our actual world is a kind of water (also called ‘heavy water’), and once we open the door to the alternative view that whether D2O is or is not a kind of water is ultimately a matter of semantic stipulation, nothing seems to prevent the Earthlings from coining a new term such as ‘dwater’ (following LaPorte's own suggestion) and drawing dwateried inductive inferences from it. Goodman's grue could become an unwelcome real possibility, given the thesis of semantic stipulation.
4.2. Four possible responses to the Goodmanian twist
(1) First response: serendipity
Given the way LaPorte designs the Deuterium Earth story, one may reply that the Goodmanian scenario is at worst innocuous, and at best a blessing. After time t (1935) the inductive generalizations that the Earthlings draw for ‘dwater’ coincide with the inductive generalizations that the Earth scientists are drawing for ‘heavy water’ as an isotopic variety of water. While gruified inferences diverge after time t, dwateried inferences serendipitously converge after time t.
But the serendipitous convergence cannot be backtrackedFootnote 45 prior to time t. Prior to time t, the same evidence (a jar of ocean-liquid) would support divergent inductive inferences for the Earthlings, who would conclude that it is not a kind of water, and the Earth scientists, who would identify it as an isotopic kind of water. Nor can one reply that prior to time t (i.e. 1935) the problem did not arise because the Earth scientists did not know about deuterium oxide, nor could draw inductive inferences about it. The Earth scientists knew about deuterium as early as 1931 when Urey identified it as an isotope of hydrogen by evaporating liquid hydrogen at low temperatures.
(2) Second response: non-projectibility
Another strategy could be to bite the bullet and concede that ‘dwater’ is non-projectible (any more than ‘grue’ is). In particular, while the vernacular kind term ‘heavy water’ is entrenched in the Earth scientists' vocabulary, ‘dwater’ is not. The former has been successfully projected in the past, while the latter has not. Moreover, the former enter in the Earth scientists' vocabulary together with a series of laws of nature; the latter does not. One could then grant that ‘dwater’ is non-projectible. But the biting-the-bullet strategy would not help here. If ‘dwater’ has to bear the semantic weight that LaPorte's story attaches to it with the thesis of semantic stipulation, it ought to be projectible for it to be an alternative live option on equal footing with ‘heavy water’.
(3) Third response: referential coincidence
Here is a more promising reply one might consider. ‘Dwater’ is in fact projectible (or, at least, as projectible as ‘heavy water’) since they are just two different vernacular kind terms referring to D2O. It is tempting to think that the Goodmanian scenario does not arise, because, after all, both prior to and after time t (i.e. 1935), ‘dwater’ picks out D2O: the Earthlings and the Earth scientists are simply using two different terms (‘dwater’ and ‘heavy water’) to refer exactly to the same liquid with the same microstructure.
Tempting as it is, there is however a major problem with this response. This line of argument is simply precluded to LaPorte, who against Putnam denies that D2O content is necessary and sufficient to fix the reference of the term. So, unless LaPorte embraces Putnam's view that D2O content is necessary and sufficient to fix the reference of both ‘heavy water’ and ‘dwater’, it is difficult to see how he could avail himself of this response. Although ‘dwater’ refers to D2O all along in LaPorte's story, precisely because LaPorte does not accept Putnam's causal theory of reference, and defends instead the view that the two terms ‘heavy water’ and ‘dwater’ pick out different descriptions about D2O (as an isotopic kind-of-water versus not-a-kind-of-water, respectively), the Goodmanian scenario cannot be easily averted. Same microstructure notwithstanding, the two terms have different intensions. ‘Dwater’ is introduced in the Earthlings' baptism to connote D2O in those aspects in which it differs significantly from ordinary water (e.g., macroscopic, chemical properties associated with its being poisonous, having different freezing and melting points). ‘Heavy water’ is introduced in the Earth scientists' baptism to connote D2O in those aspects in which it bears sufficient similarity with ordinary water (e.g. microscopic, physical properties associated with its isotopic nature). ‘Dwater’ connotes D2O as not-a-kind-of-water. ‘Heavy water’ connotes D2O as a-kind-of-water. By connoting different properties of D2O (or, if you like, by associating different descriptions, or different intensions to D2O), the two terms are opaque under co-extensive predicate substitution, with the result that they cannot be interchanged in inductive inferences salva projectibility: while ‘heavy water’ is projectible, ‘dwater’ is not.
In particular, in 1935, upon their return to planet Earth and upon hearing the above report from the Earth scientists (as per Section 3), the Earthlings are likely to accept that from a scientific point of view the paradigm sample of Deuterium Earth's ocean liquid bears the same (or sufficiently similar) microstructural kind relation to what we ordinarily call ‘water’, making ‘dwater’ no more projectible than ‘grue’.
(4) Fourth response: stipulation versus discovery
One might rejoin that LaPorte's view escapes the problem because we did not discover that D2O is a kind of water any more than we discovered that H2O is water. If it is just a matter of semantic stipulation whether D2O counts as a kind of water, it is not after all paradoxical that something that does not count as a variety of water prior to time t, may count as such after t.
But the problem remains. For the case to be made that the Earthlings' option of not counting D2O as a kind of water is semantically on a par with the Earth scientists' alternative option (hence, for the two terms ‘dwater’ and ‘heavy water’ to be interchangeable in inductive inferences), ‘dwater’ would have to be construed as synonymous with ‘heavy water’, i.e. it would have to be surreptitiously construed as connoting D2O as a kind-of-water. It is not surprising that ‘coincidentally’Footnote 46 (and in blissful ignorance of) the Earthlings arrive at the same conclusions of the Earth scientists about the atomic microstructure of deuterium oxide (without in fact even knowing anything about isotopes, or knowing the difference between microscopic atomic-number-dependent and macroscopic atomic-weight-dependent properties), and even about its chemical reactions in a H-bomb, twenty or so years before one was in fact built and tested on planet Earth (in the early 1950s with the consequent destruction of the Bikini Atoll in Micronesia). Without mentioning that the ‘sophisticated techniques’ through which the Earthlings eventually learn about the unusual hydrogen composition of the ocean-liquid are presumably one and the same distillation techniques used by Harold Urey in 1931 to identify the hydrogen isotope.
In sum, to escape the Goodmanian scenario and argue for the projectibility of ‘dwater’ as a semantically live option on a par with ‘heavy water’, one must pack into the term ‘dwater’ as many scientific assumptions as possible about the D2O content as an isotopic kind-of-water, including its far-away-in-the-future performance in an H-bomb (with a Deuterium Earth-equivalent of the Bikini Atoll). In this way, the projectibility of ‘dwater’ becomes parasitic on the projectibility of ‘heavy water’ as an isotopic kind of water.
But if the projectibility of ‘dwater’ is ultimately parasitic upon the projectibility of ‘heavy water’ (because scientific facts about the D2O content determine the reference of ‘dwater’ as much as they determine the reference of ‘heavy water’), then the two are no longer alternative options. The Deuterium Earth story would no longer support the thesis of semantic stipulation, according to which we could have gone either way in taking D2O as a kind of water or not.
5. Conclusion
This is after all only a story. We do not have competing chemical schools, classifying D2O, respectively, as a kind of water or not (by contrast with biology, where there are competing biological schools for the classification of higher taxa). Perhaps an account of what makes both ‘heavy water’ and ‘dwater’ projectible (without falling back onto Putnamian assumptions about the D2O content being necessary and sufficient to fix the reference of the terms) is available within the resources of LaPorte's view. Perhaps rigidity de jure of kind terms can be of some help here. But LaPorte gives no indication of how an account of projectibility would look like within the resources of his view of natural kind terms. And without such an account, the received chemical orthodoxy that D2O is an isotopic kind of water, and that Deuterium Earth's ocean is watery after all, seems to stand unscathed.
