Brain Death and Disorders of Consciousness, Edited by C. Machado and D. A. Shewmon Kluwer Academic/Plenum Publishers, New York 2004
THE SEMIOTICS OF DEATH
AND ITS MEDICAL IMPLICATIONS
D. Alan Shewmon and Elisabeth Seitz Shewmon*
Die Grenzen meiner Sprache bedeuten die Grenzen meiner Welt.
[The limits of my language signify the limits of my world.]
(Ludwig Wittgenstein, Tractatus Logico-Philosophicus 5.6)
In any discussion pertaining to this world, human beings are bound to the medium of language. We are often unaware of the degree to which language, as the physical shape of all thought, exerts its influence on the distinctions we make and consequently on the formation of our notions and ideas. The way we think shapes the way we speak, but also conversely, the language we speak shapes the thoughts we think. The languages we live in are the result of all preceding evolution of thought, cultural interaction and the communication thereof within a linguistic community, already presenting us with a pre-formed way of structuring and interpreting the world (Weltanschauung) at the time of primary language acquisition. We are not able to step out of the medium of language, but we are able to step out of one particular language into another, thereby critically evaluating the distinctions, terms and notions we usually take for granted. Surprising new perspectives open up on seemingly well-known objects of debate, and from the meta-linguistic point of view, more often than not what appeared to be a problem on the object level turns out to be an inherent feature of the language we use to discuss it. A case in point is the phenomenon called death.
1. INTRODUCTION
This presentation will (1) elaborate on the ambiguity of the concept corresponding to the word “death” and (2) suggest that the “dead donor rule” has caused us to focus all along on the wrong ethical question surrounding transplantation. The first part will be
* D. Alan Shewmon, MD, Professor of Neurology and Pediatrics, David Geffen School of Medicine at UCLA,
89 D. A. SHEWMON AND E. S. SHEWMON 90
structured according to three levels of linguistic consideration as they relate to the death debate: (1) respecting word-usage conventions for the sake of communication (and the ill-formulated process-vs.-event debate), (2) the role of vocabulary in reflecting importance and focusing attention (and the multiplicity of death-related events), and (3) the dynamic interaction between language and thought (how much might our unitary concept of death be conditioned by the language we think and communicate in). The final section will consider the implications of all this for transplantation.
But first let us recall the traditional hierarchical schema (concept-criterion-tests) for discussing the definition and diagnosis of death and the associated assumption that “beginning at the beginning” means beginning with a concept of death. This structured approach was logical and served the field well for many years; nevertheless, the best point of intellectual departure may actually not be a concept, but rather observation – not only of various death-scenarios but also of how people speak and think about death.
1.1. Concept-Criterion-Tests: Sounds Nice but Gets Us Nowhere
In the four-decade-old debate over the definition and diagnosis of death, virtually everyone has accepted the conceptual framework eloquently outlined by Bernat and colleagues in their seminal paper of 1981,1 namely (1) to distinguish clearly between three levels of discussion – concept, criterion, and tests – and (2) to respect the flow of logic in just that direction: from the abstract to the concrete. It has seemed obvious to most participants in the debate that the only way to get a proper handle on “death” is to begin with a correct (or at least agreed upon) definition or concept of death (a purely philosophical matter), then determine what anatomical criterion instantiates that concept (a hybrid philosophical/medical matter), and finally determine what clinical signs or tests reliably indicate that the criterion is fulfilled in a concrete case (a purely medical matter). After all, how can one devise tests for an unstated criterion, or how can one formulate a criterion for an undefined or vague concept?
For many years the first author also accepted this modus operandi as the only rational way to attack the problem of diagnosing death,2 but he is no longer so sure. First of all, there are reasons to doubt whether a valid “concept of death” can even be clearly formulated. It is inextricably bound up with the concept of “life”: death is the cessation of life. In his book “What is Life?” philosopher Josef Seifert convincingly argues that “life” is intrinsically undefinable.3 It is an “ur-phenomenon,” conceptually fundamental in its class; no more basic concepts exist to which it can be reduced. It can only be intuited from our experience of it, both around us (in nature and human society) and especially within us (our own personal psycho-physical life). Not even the thermodynamical concept of life as localized anti-entropy,4-7 which appeals especially to scientists, adequately captures the essence of life in general. To cite one of Seifert’s examples, a spore can remain inert for centuries but still possess life. If there can be no totally adequate definition of “life,” neither can we presume to begin the logical cascade of concept-criterion-tests with a totally adequate “concept of death.”
Many commentators have pointed out that, depending on the extent to which a concept relates to the physical domain, it may or may not specify a physically definite moment, process or state, and hence be correspondingly more or less relevant to the clinical diagnosis of death. For example, “cessation of the organism as a whole” is both a
SEMIOTICS OF DEATH 91
concept of death and the presumed description of a physical event (even if unobservable directly). (The precision with which it can be pinpointed in practice is a separate matter.) Some philosophical concepts of “soul” (e.g., the Aristotelian-Thomistic “substantial form of the body” officially endorsed by the Catholic Church) correspond to “organism as a whole” in the biological domain; thus, “departure of the soul” (so understood) corresponds in principle to a definite, physically definable moment. Other concepts of soul, however, (e.g., in Platonism and eastern religions) do not imply anything in the physical domain related to its coming or going. What becomes of the three-level schema if a philosophically favored “concept of death” happens to imply no clear empirical consequences for biology or medicine?
In some socio-medical contexts the very “concept” of death is what the three-level schema would call a “test.” For example, in an anticipated home death, when the family at the bedside relates the onset of death to final apnea, their implicit concept of death is “expiration,” quite different from, but no less appropriate than, “loss of the organism as a whole,” “irreversible loss of the capacity for consciousness,” “loss of potential for cardiac autoresuscitation,” or any of the other popular candidate concepts in the current medical-ethical debate. The “concept” of death is similarly at the level of “tests” when a physician declares (again, appropriately) the official time of death of a monitored, non-donor patient with a do-not-resuscitate order to be when the EKG goes flat – although one could argue that this is not really a concept (but what more genuine concept is it a “test” for?) or that it is merely a legal fiction.
Veatch rightly distinguishes two basic types of concept of death: normative and ontological.8 Those who favor a unitary, objective concept tend to think of death ontologically, whereas those who favor a context-dependent notion tend to think of it normatively. We submit that a normative concept implies a tacit ontological concept: if death is defined as the moment after which, or state when, it is ethical to treat the person as no longer alive, there is an already understood ontological concept of death. Moreover, in what follows we shall suggest (contrary to the first author’s previous writings) that there is no singular, clear ontological concept of death, as the single word “death” tempts us to imagine, but rather a legitimate context-dependency even in the ontological order.
Therefore, instead of proceeding via logical deduction from some preconceived, abstract “concept of death,” we propose a phenomenological approach: to step back from linguistically constrained notions (to the extent possible) and simply look afresh at the spectrum of realities surrounding what people call and have called “death” – as though one had never heard of the term “death” and feels no compulsion to label any particular phenomenon “death.”
This mental exercise could prove invaluable and liberating, if in fact it is a delusion that our concepts of life and death are simply generalized from everyday experience and labeled by linguistic convention with the words “life” and “death,” when in actuality those very concepts have been shaped by our language, and more importantly constrained by it (particularly by the poverty of a lexicon that offers only the single dichotomy “life” vs. “death” to apply to a rich and complex portion of reality). If our language deceives us into thinking that there must be only one reality of death merely because we have the one word “death,” no wonder that we end up arguing endlessly and uselessly over which of the various events or processes is the real “death.” No wonder also that the death-debate is so full of logical inconsistencies and failures to communicate, because people use the same word to express different concepts, not quite realizing (precisely because of the single word) that the concepts are different.
2. WATCH YOUR LANGUAGE!
THE POINTLESS DEBATE OVER PROCESS VS. EVENT
Even the most cursory re-examination of death-related phenomena reveals the existence of various processes and events, all of which are equally “real.” The debate whether death is a process or an event has dragged on for over three decades since the famous kick-off pair of articles by Morison9 and Kass.10 Nothing substantially new is ever said, and proponents of each side simply talk past each other. As traditionally formulated, the debate is intrinsically unresolvable, because it is based on terminology that is linguistically incorrect and fails to adequately convey the real issue.
A variation on the “process” theme is Linda Emanuel’s “asymptotic model,” which rejects a dichotomy between life and death and focuses instead on dying as “a bounded zone of residual states of life.”11 One problem with the asymptotic metaphor is that by definition an asymptote never reaches its limit. That means that even with rigor mortis and putrefaction, the corpse may be dead “for all intents and purposes” but still isn’t finally and totally dead (and never will be). If the approaching curve actually reached its limit, it would not be properly speaking an “asymptote”; moreover, there would then be an “event” that could legitimately be called “death” or “the moment of death.” A more serious problem with the model is the axis labeled “life” or the multiple axes proposed for the expanded metaphor (“personhood” and “biological life,” each of which could in turn be divided into various sub-axes: “self-awareness,” “ability to love,” etc., at the “organic, cellular and molecular” levels). To treat such things as quantitative parameters representable as orthogonal axes in multidimensional space is more poetic than scientific. What could possibly be meant for any such axis by a quantitative value along the continuum of real numbers, and how would such a value be determined? If “life,” “personhood,” etc. are not meaningfully quantifiable, neither is the asymptote a helpful mathematical metaphor.
2.1. Precedents for State-Discontinuity Brought About by Continuous Change in Parameters
Process advocates have universally failed to comprehend (and event advocates have failed to point out) that invisible though real discontinuities of state can result from continuous changes in observable parameters. Continuity and discontinuity are mutually contradictory only in the same domain; system-state and parameter values are in different domains. To deny state discontinuities simply on the grounds of continuity of underlying processes is unscientific and anachronistic; it ignores some of the most interesting developments in modern science and biology. [1]2.1.1. Quantum Mechanics and Eigenstates
Quantum mechanics has something important to teach in this regard. Take the most elementary quantum mechanical system, a hydrogen atom. There is an infinite continuum of possible positions in space where the electron might be found, but only discrete possible energy levels and corresponding probability distributions of position (specified by the Schrödinger equation). No matter where the electron is in space, its position is mathematically compatible with either a free or a bound state, and if bound, with any of an infinite number of discrete energy levels. Consider a free electron shooting towards a free proton; suddenly a quantum of energy is emitted and we now have a hydrogen atom. Despite the continuity of the electron’s trajectory, a dramatic yet unobservable event takes place: one moment there are two free particles and an infinitesimal moment later there is one atom.
If this is the case for something as simple as a hydrogen atom, how much more should we expect to find state discontinuities in more complex quantum systems and even in macroscopic systems whose differential equations, like the Schrödinger equation, yield discrete eigenvalues (the mathematical basis for the “quanta” of quantum mechanics). Varela made a good case that the dynamics of living systems must be mathematically modeled by enormous, complex sets of differential equations whose solution sets, as in quantum mechanics, contain discrete eigenvalues, corresponding to discrete eigenstates.12 When a continuous change in one or more parameters occasions a change from one state to another, the state-change is necessarily discontinuous (like the transition between energy levels of a hydrogen atom). If that is the case within life, it should hardly be surprising or incomprehensible at the transition between life and non-life.
2.1.2. Nonlinear Dynamics and Bifurcations
The burgeoning field of nonlinear dynamics makes the process-event debate seem almost antediluvian. Apart from eigensolutions, nonlinear differential equations often occasion a type of state-discontinuity known as bifurcation. A classical example is the Hopf bifurcation, which describes the behavior of an oscillator subjected to a particular pattern of external influence. It can be represented by the following pair of equations:13
dx/dt = – y + x {μ – (x2+y2)}
dy/dt = + x + y {μ – (x2+y2)}
The qualitative behavior of the system is determined by the control parameter μ. For μ<0, the system undergoes a damped oscillation asymptotically approaching an equilibrium point. For μ>0, however, the system asymptotically approaches a “limit cycle” of self-sustained, spontaneous oscillation with radius √μ and period 2π. As μ is varied continuously from negative to positive, when it passes the bifurcation value μ=0, a qualitative state-change occurs: an infinitesimal instant earlier the system tended to static equilibrium, and an infinitesimal instant later it tends to spontaneous oscillation. This radical discontinuity in system property is very real, though at the moment quite D. A. SHEWMON AND E. S. SHEWMON 94
unnoticeable by anyone merely watching the oscillator and following the continuously changing values of x, y, and μ.
Bifurcations of this and other sorts are ubiquitous throughout nature, especially biological nature.14-18 The transition between homeostasis (characterizing life) and lack thereof (characterizing decay) is a bifurcation not unlike the Hopf bifurcation, except on an infinitely more complex scale.
2.2. Linguistic Respect
The whole process-event question is simply mis-posed as necessarily “either-or.” Between the beginning of dying and the end of decaying, continuous processes and discontinuous events occur (some directly observable, others not). Whether we assign the word “death” to one or more of these processes or events is primarily a linguistic question, not a biological one.
Death has always been understood as the cessation of life. But within this general notion the word “death,” like many words, can have more than one nuance of meaning in common parlance, or in Saussurian terms, a range of signifiés. It can refer to (1) the state of the remains of a previously living thing, (2) the state of the remains of a previously physically living thing, which still lives in some spiritual sense, or (3) the moment of transition from living to one of these two states. “State” and “moment” are not opposed concepts of death but merely acceptable alternative usages of the word “death.” When we employ it in the third sense, e.g., “death occurred at 2:15,” this is merely another way of saying “[the state of] death began at 2:15.” By contrast, the adjectival forms “alive” and “dead” apply only to states.
The referent of the verb “to die,” however, depends on its form. The simple present or past tense is usually understood as referring to the momentary act of transition from life to death, e.g.: “He died at 2:15,” or (last words of a martyr) “I die for the cause of justice!” (referring to the immediate future). The present perfect, “has died,” indicates a present state of death that began at the moment of death. The progressive form, present participle, and gerund indicate, by their very linguistic nature, a process: “dying,” “is dying,” etc. can only refer to the gradual fading of life processes, tending toward [the moment of] death (= ending with the beginning of [the state of] death). (For an extensive analysis of the aspectual qualities of “die” verbs in 18 languages, see Botne.19)
We therefore fail to see the novelty of James DuBois’s proposal of death as a “state,” as opposed to an event or process.20 Of course it is a state, which begins with an instantaneous transition-event that linguistic convention allows us also to call by the same name “death.” What is provocative about DuBois’s proposal is rather that the state of death need not, by definition, be irreversible – but more on that below.
There is nothing particularly esoteric about all of this. There are processes and events related to death, and the English language provides us with the means of communicating accurately about it all and being understood. Nothing is to be gained, and only confusion sown, by insisting that a word which on linguistic grounds refers to a moment or a state must henceforth be used to refer to a process.
Some antithetical pairs logically entail a continuum of gradations, such as light and dark, rich and poor, beautiful and ugly. Any transition between the two is necessarily gradual; there is no intelligible, non-arbitrary point along the spectrum of illumination-SEMIOTICS OF DEATH 95
intensity at which “light” becomes “dark.” By contrast, there can be no intelligible continuum between logical contradictories, such as pregnant and not-pregnant. Properties traditionally associated with life and death form such pairs: being vs. non-being, unity vs. multiplicity, endogenous opposition to entropy vs. tendency toward increasing entropy, self-sustained limit-cycle vs. damped oscillation tending to static equilibrium (in the Hopf bifurcation analogy). Between any of these and its opposite the transition must, by logical necessity, be instantaneous and discontinuous, as must also be the case with the intimately related dichotomy life vs. non-life (i.e., death).
It would be linguistically possible, though uninsightful, to claim that “life” and “death” have been misconceived all along, that there are only graded realities like “light” and “dark,” and that there is no discrete event or moment that separates the two, only a fuzzy transition. But it is simply an abuse of language – which biologists, physicians and ethicists have no right to change – to insist that “death” is a “process” when one could just as well use the term “dying” instead.
This foolish debate may be partly language-specific. Among death-related expressions in German, for example, there is the noun Tod (death) and the verb sterben (to die), with its nominalized form das Sterben, referring to the act or process of dying. The difference between the two nouns is clearly reflected when they are juxtaposed in a title, such as “Tod und Sterben” (corresponding to the English “On Death and Dying”). The debate whether death is an event or a process would already for that reason not make sense in German, because if one wants to speak of a process one uses Sterben, and if one wants to speak of an event one uses Tod. It would be incomprehensible even to ask whether Tod is a process or Sterben an event. Since English has only the one noun “death,” and since the nominalized form of “to die” derives etymologically from the same root, it is easier to fall into wondering unwittingly whether “death” is a process or an event, but the question is still just as malformulated linguistically as it would be in German regarding Tod.
3. MARK MY WORD: A VOCABULARY OF DEATH-EVENTS
The lexicon of any culture or society reflects what is important for them; conversely, children learn what distinctions are important in the world around them partly by the language they learn. Pilots and weather forecasters have many terms for what the general public calls “clouds.” Their specialized vocabulary is not merely descriptive; it is essential to the skilled and safe performance of their task. Sylvan societies have an expanded vocabulary for plants, serving not only a communicative function (reflecting the interests of adults) but also a pedagogical one (helping children, through learning the names, to learn the distinctions signified by the names: what is safe to eat, how to make various medicines and paints).
Eskimos and skiers have a multiplicity of words for frozen water. One reference on West Greenlandic, for example, identifies at least 49 different words/lexemes referring to ice and snow.21(pp.366-367) The linguistic controversies over the exact number, the distinction between “word” and “lexeme,” etc. do not concern us here. The point is that what to us are incidentally different qualities or locations of “snow” (snow on the ground, snow falling in air, air thick with snow, feathery clumps of falling snow, newly D. A. SHEWMON AND E. S. SHEWMON 96
fallen snow…) are perceived and conceived by the natives of western
Now imagine that pioneers from the tropics migrated to the arctic and suddenly had to learn how to survive there. Their very language, which allows them to see only “snow” and “ice” all around, is itself detrimental to survival. They would do well to learn the Eskimo vocabulary in order to focus attention on important aspects of their new reality, critical distinctions regarding that white stuff they so carelessly refer to as “snow.”
This is precisely the situation we find ourselves in regarding “death.” We have migrated through human history into the modern ICU, bringing with us the linguistic baggage of a relatively simple concept of death for which the one word had always sufficed. Now we find ourselves in a situation for which medically real and ethically critical distinctions lack words in the common vocabulary. The best we can do is to speak in awkward paraphrases, such as “the point in time beyond which cardiac auto-resuscitation is impossible.” To ask which of these technological mumbo-jumbos is really “death” may perhaps be as linguistically and epistemologically inappropriate as asking an Eskimo which of sullarniq, aput, qaniit, nittaalaq, ... is really “snow.”
Having established above that system-level events can and must occur along the process of dying/decaying, and following a phenomenological approach, let us simply examine some sample events, each in its own proper context, putting aside the question as to which is the one “true” death. (Why must we assume a priori that there is only one?) Because our language lacks simple names for the technologically defined events, let us try to free our minds to the extent possible by simply calling the candidate events E1, E2, and so on, as defined below. (Such neutralized nomenclature is similar in approach to what Lynn and Cranford have labeled “Time 1,” “Time 2,” etc.22 Some events are directly observable, such as E1 and E2; others can only be inferred as having happened already or not yet. Depending on the type of event, such inference could be based on something observable (e.g., with E5, inferring unconsciousness from unresponsiveness) or on cumulative knowledge from scientific studies (e.g., E3, E4, E6, E7).
Let us begin with common notions of death as an event. It could be instructive to recall what death-scenes were like before ventilators made possible a dissociation of the chief physiological components and before transplantation became a utilitarian motive for changing traditional concepts of death.23,24(pp.20-37),25 Given how foreign an environment modern ICUs are to most people, perhaps the average non-medical person’s understanding of death might not be that different today from what it was in pre-ICU times, once we get beyond the uncomprehending parroting of certain “brain-death” rhetoric picked up from the media. The first two E’s to be considered are of this type.
3.1. E1 – Expiration (“Ex-Spiration”)
A common pre-ICU-era death scenario was the terminally ill person lying in bed at home, surrounded by family members keeping vigil. Breathing becomes shallower and SEMIOTICS OF DEATH 97
finally stops. As soon as the onlookers realize that the rhythm of breathing has halted and that a next breath is not going to come, they infer that with the end of that final exhalation the loved one died. Someone closes the deceased’s eyes (if not already closed) and the family begins mourning. There is no checking for pulses or EKG electrical activity or brain waves. In such a context final apnea constitutes the most glaring (perhaps the only evident) discontinuity between the process of dying and the process of decaying. Let us call this event E1.
Of course, as Emanuel points out, actual death scenarios aren’t always so straight-forward.11 There may be irregularly timed agonal gasps. What the family thinks is the last may turn out not to be so. Such a pattern of terminal respiration may temporarily complicate matters for the family in deciding when E1 occurs, but it does not undermine the concept of E1. No matter how many times the family might be fooled into thinking that a certain breath was the last, eventually there will be a truly last breath, which will in retrospect be known as such with certainty. The end of that final exhalation is E1.
It was common to declare, in colloquial theological terms, that at this moment the person “gave up the ghost.” In Jewish tradition, cessation of breathing has generally been considered a more critical and essential marker of death than cessation of circulation.26(pp.241-254),27,28 One of the Yiddish expressions for “to die” is oyshoykhn di neshome, in which neshome means “soul” and oyshoykhn means “breathe out” (i.e., one’s final breath).29(p.222)
Christian tradition likewise contains an E1-precedent for the moment of death in the narration of Jesus’s own death on the cross. All four gospels link his final crying out with the giving up of his spirit and illustrate the close relationship between breath and soul in English, Latin and Greek (Mt 27:50, Mk 15:37, Lk 23:46, Jn 19:30). In modern English, “to expire,” meaning “to die,” is etymologically derived from the Latin exspirare, “to breathe out.” Further instances of the conceptual connection between “life” and “breath” are the English words “(to) animate” and “animation,” which derive from the Latin anima or animus, “soul” or “spirit.” In ancient Greek these correspond to two words, each of which is related to both life and breath: Psychê (ψυχή) is translated most often as life or soul, but also as ghost, departed spirit, conscious self, personality; whereas the verb form means “to breathe,” “to blow.” Pneuma () signifies “breath,” “spirit,” “mind”; the verb “to blow,” “to breathe out.” The perfect means “to possess a soul,” “to be animated,” “to be alive,” and “to have understanding”; whereas is a favorable wind.
ψύχωπνεῦμαπνέωπέπνυμαιἡπνέουσαOf course we know from modern medicine that in scenarios where death is understood as apnea (or at least coinciding with apnea), had the pulse been measured, it would have been noted to persist some tens of seconds after the final exhalation. Ineffectual EKG electrical activity might persist some minutes afterwards; all brain function would not be irreversibly lost for some minutes more – and who knows how long after that the spiritual soul really departs to enter the afterlife. In his previous unitary-death mind-set, the first author would have said that E1 was a convenient fiction for non-medical people in a non-medical setting, and that the “giving up of the ghost” at that moment was forgivable pious poetry, but that the “real” moment of death occurred some time later (whether defined philosophically in terms of loss of the substantial form of the body, or systems-dynamically in terms of loss of integrative unity of the organism D. A. SHEWMON AND E. S. SHEWMON 98
as a whole, or thermodynamically as the transition from endogenous opposition to entropy to a giving-in to entropy – all of which were assumed to necessarily coincide, because they were assumed to be merely different ways of looking at the same phenomenon).
But does our medical and philosophical understanding mean that we should intrude on the family’s death-bed scene to explain that they ought not to begin grieving yet, because their loved one isn’t really dead yet, and won’t be for some unknown number of minutes more? Does the absurdity of such an imaginary intrusion derive simply from its impoliteness and insensitivity, or rather from some deeply seated intuition that perhaps the family was right all along to understand that their loved one really did die and “give up the ghost” with that final exhalation? Why must an ICU context be the only “correct” one for understanding death?
Some authors have argued that irreversibility is not, and has never been, intrinsic to the notion of death.20,30 Resurrection stories in religious history may be miraculous or mythical, but they do not do linguistic violence to the concept of death. Citing some fascinating studies on near death experiences, DuBois suggests that it may be reasonable after all to regard such patients as transiently, reversibly in the state of death, and that the European term “reanimation” might be literally more correct than our “resuscitation” after all.20 The German term Wiederbelebung expresses the idea even more explicitly (literally “the giving back of life” to somebody). DuBois would not declare the state of death until all three major bodily systems have shut down (at first reversibly, soon irreversibly): breathing, circulation, and brain function.20 But why must even this be so?
Let us take his proposal one step further. In the kind of scenario we are talking about, the person “falls asleep” (a common poetic expression for “dies”) with the last breath: consciousness is lost, even if all brain functions are not yet lost. Why should a residual heartbeat for the next half-minute or so preclude this state from being death already, any more than it should preclude “brain death” from being death? For those who accept the notion of soul as life-principle, why could it not be that, rather than asystole causing or occasioning the soul to depart, the heart ends up stopping precisely because the soul has already departed? If hypothetically the person were successfully “reanimated” by doctors, this would not prove that the soul had not yet departed and that the person was not really dead; it could merely indicate that the soul had returned.
The more one gets into the role of the devil’s advocate about all this, the more it becomes obvious how, in the home death-bed context, the various physiologically and systems-dynamically defined “events” are relatively abstract, hypothetical, unobservable, and without the least practical consequence, compared to that dramatic final exhalation. Who are we to say that the person didn’t really die then and that the family was “wrong” (even if forgivably so) to initiate death-behaviors at that moment? Would it not rather be like forcing a square peg through a round hole to insist that one of the several candidate ICU-understandings of death be applied to this context?
3.2. E2 – Final Asystole
Nowadays the death-bed scene often takes place in a hospital, with the patient connected to various monitors. Suppose a do-not-resuscitate order is in place for a terminally ill patient. At one point the person stops breathing, and everyone watches the SEMIOTICS OF DEATH 99
cardiac monitor until the rate slows, the pulse weakens, and finally the EKG goes flat. Let us call this latter event E2. At that moment the physician or nurse notes the official “time of death” to be entered on the death certificate, and the family begins its grieving process. (If the scenario were different such that E2 preceded E1, as in sudden cardiac arrest, it would of course be more appropriate to call E1 death.)
It might happen that some other pulseless cardiac rhythm preceded asystole, such as ventricular fibrillation or electromechanical dissociation. It is a moot point whether the “official” moment of death for purposes of the death certificate and pronouncement to the family is at the onset of such rhythm or of asystole. If desired, we could subdivide E2 into two cardiac-related events: the onset of pulselessness (E2a) and the onset of asystole (E2b); if bradycardia slows to asystole without any other terminal arrhythmias intervening, then E2a and E2b will coincide.
Are staff and family “wrong” in considering E2 the moment of death, given that, in principle, for the next few minutes resuscitative efforts could succeed in restoring cardiac function? If the state of death need not be intrinsically irreversible, as DuBois suggests, the answer is clearly “no.” Or would they be “wrong” in the other direction, given that, had the same terminal illness ended at home unmonitored, they would have identified E1 as the moment of death? We do not think so. Since each “moment of death” seems intuitively appropriate within its own context, why must one trump all but one with the assumption of a unitary concept, as though correctness of E1 implied erroneousness of E2 and vice versa? Why not rather give priority to the intuitive correctness of both over the unitary-concept assumption?
3.3. E3 – Loss of Potential for Cardiac Auto-Resuscitation
The moment of loss of potential for cardiac auto-resuscitation (let’s call it E3) seems to be of interest only in the context of non-heart-beating organ donors (NHBD). Some critics of the
3.4. E4 – Loss of Potential for Interventional Resuscitation
The NHBD literature abounds with debate over the various senses of “irreversibility” of apnea and asystole. The “ethical” interpretation (“irreversible” D. A. SHEWMON AND E. S. SHEWMON 100
because one should not and will not try to reverse it) corresponds to E1-E3,32 whereas the “physiological” interpretation (loss of physical potential for resuscitation despite all technological means) corresponds to E4. Even E4 can be subdivided into “strong” and “weak” interpretations of irreversibility (“not reversible now with existing technology” vs. “can never be reversed despite all future technology”).30,33 Unlike E1 and E2, which are distinctly observable events, E3 and E4 are unobservable (based on subtle molecular-level changes), but no less real and distinct, moments along the continuum of physiological changes following hypoxic-ischemic damage to the heart.
3.5. E5 – Onset of Permanent Loss of Consciousness
This is the analog of E2 in the neurological domain, occurring around 5 to 10 seconds or so after (permanent) cessation of blood flow to the brain. If there is no moral obligation to try to restore consciousness (e.g., through restoring circulation of oxygenated blood), and especially if there is an obligation not to try, then E5 is also analogous to E3. Objections to NHBD on the basis that the donors are not yet “brain dead,” i.e., that they are at E5 but not yet E6 or E7 (vide infra), are analogous to the cardiac-domain objections that donors are at E3 but not yet E4. Both domains of objection focus on the wrong question (“Are they dead?” vs. “Will they be killed or harmed?” – see final section below) and gratuitously absolutize one particular E for no apparent reason other than to build a case against NHBD.
3.6. E6 – Loss of Potential for Recovery of Consciousness
E6 corresponds to the notion of “irreversible coma” or (in the case of damage primarily to the cerebral hemispheres with a relatively intact brain stem) “permanent vegetative state.” As with the cardiac E3 and E4, our ability or inability to determine in practice when E6 occurs, to such-and-such degree of precision, is an important empirical issue but is beside the point of the present conceptual discussion. If what sets the death-events in motion is primarily neurological rather than cardio-respiratory (i.e., all the sorts of etiologies that typically lead to “brain death” or persistent vegetative state), then the sequencing of E5-E7 relative to E2-E4 will be reversed, and the sequencing relative to E1 will depend on the precise temporal-anatomical pattern of brain insult.
3.7. E7 – Irreversible Loss of All Brain Function
This is how the Universal Determination of Death Act and most US state laws formulate the neurological arm of the bifurcated statutory definition of death, a condition commonly but unfortunately6 known as “brain death.” The fact that present diagnostic standards do not actually require strict fulfillment of this statutory definition (by allowing for some residual brain functions dismissed as “insignificant” or not true “functions” – cf. Shewmon7,34) is beside the point that E7 is a real moment along the continuum of death-related physiological changes. If desired, we could subdivide E7 according to the various schools of “brain-death” advocacy, such as E7a for the British “brainstem death” and E7b, E7c, … for irreversible loss of all “critical” brain functions, depending on how one chooses to define “critical”.35,36 SEMIOTICS OF DEATH 101
3.8. E8, E9, …
We could think of many other potentially definable, real physiological events along the course of dying/decaying and give each an E-label. Who knows whether we would come up with as many clinically or morally relevant ones as the 49 types of ice and snow? For purposes of this paper, the principle has been sufficiently illustrated with E1-E7. For each of these events we could coin a name more descriptive than “E3” and less of a mouthful than “loss of potential for spontaneous cardiac auto-resuscitation,” just as the West Greenlanders say “siirsinniq” instead of “S45” or “ice swelling over partially frozen river, etc. from water seeping up to the surface.” But that exercise can be saved for a more propitious occasion.
3.9. Special Contexts
In certain rare contexts, for example decapitation, some of these E’s may not apply and other E’s become uniquely relevant, such as the moment of head-severing. In other rare contexts, such as instant vaporization from a nuclear blast, all the E’s occur simultaneously. Although of possible theoretical interest, such unusual death-circumstances have little relevance to the “dead donor rule,” the “physiological decapitation” analogy of “brain death” notwithstanding.
4. WE THINK WHAT WE SAY: IMPLICATIONS FOR A UNITARY CONCEPT OF DEATH
The dynamic interaction between language and thought goes much deeper than merely focusing attention by naming. What if our assumption that there must be a clear, unitary, objective, correct concept of death, is derived not so much from intellectual insight as from an accident of the language we think in: the singularity of the word “death”? What if our very lexicon is a set-up for the interminable and seemingly unresolvable debates about the nature and determination of death, as well as for the incoherent thinking about death that abounds among not only the general public but health professionals as well?
In the following, we will briefly survey some selected languages and their lexical material referring to the phenomenon of “death” and “dying”. For this purpose, the authors’ own linguistic competence was combined with a little field study among native speakers and an inquiry addressed to members of the Linguist List, an internationally acknowledged online forum for linguistic discussion. Naturally, a systematic study investigating all semantic and stylistic aspects involved would be desirable, but beyond the scope of this paper.
Most languages possess a single-word equivalent to the English “death,” suggesting that there is indeed a corresponding singular concept or “ur-phenomenon” universally understood across societies down through history. This makes sense, because up until the very recent advent of “life support” in developed countries, the set of candidate death-events was fairly limited (final breath decapitation…). Moreover, nothing critical hinged D. A. SHEWMON AND E. S. SHEWMON 102
on the exact timing of death (so long as it had surely occurred prior to burial). But modern developed countries now find themselves with death-situations unknown and inconceivable throughout the millennia over which languages developed. Therefore, just because we grew up learning to speak and think with the one word “death,” it does not follow that we must also think with the same singular concept in the context of modern ICUs. (Neither does the new context necessarily imply that we shouldn’t think in terms of a singular death-concept; it simply raises the question, which we believe is answered in the course of this paper.)
Wilhelm von Humboldt (1767-1835), one of the “fathers” of modern linguistics, developed as an integral part of his work a theory of the dynamic mutual interaction between how we think and how we speak. We use language to express ideas, but our range of ideas is also shaped and limited by the language we grow up in as we learn to think. This limitation can, of course, be transcended by reflecting about the preconditions provided by our native languages. Some of the more spectacular examples of the degree to which languages can differ structurally are the Algonquin language, which has no word for time,37 and Classical Hebrew, which operates without the category of tense (for the tense/aspect discussion, see Cook38), as do Chinese languages. Although it is not true that speakers of these languages do not understand the concept of time, it has to be inferred that such structural details of a language do shape the world view of its speakers.
Distinctions in language derive from the awareness of distinctions in reality; but the reverse is also true: the perception of distinctions in the real world is facilitated when a linguistic distinction already exists. Speakers of languages with a reduced color lexicon, for example, consider as mere variations in a single color what others regard as fundamentally different colors.39 In Russian there are two different colors, both called “blue” by English speakers (sinij for dark blue and goluboj for light blue), and there is no word corresponding to our genus “blue.”
To what extent does this phenomenon influence our thinking about death? In other words, do we have the single word “death” to express a primarily intuited, unitary concept of death, or is our concept of death unitary because up to now we have had only the one word for it (apart from pejorative or euphemistic synonyms)? By examining death-words in other languages and cultures, we might get a sense of the extent to which the (largely English-speaking) debate about death might be conditioned or constrained by the English language.
Some languages have no equivalent for the English word “death.” For example, in the Kovai language of
A similar thing occurs in Quechua:
“My sister-in-law is dying!” This, in Quichua, may mean anything from a headache to a snakebite. If one is in excellent health, he is “living.” Otherwise, he is “dying.” “What is the matter with your sister-in-law?” “She is causing a child to be born. Will you come?”40(pp.42-43)
Such a linguistic difference reflects a profound difference in world-view, in which death is viewed not as the end of life but as a kind of extreme of illness, after which the spirits of the dead continue to live (physically) in a different place, eating, sleeping, working, etc., from whence they may return periodically to speak about their present life to family members in dreams.
A parallel phenomenon occurs in Nivkh (older name: Gilyak; spoken on
The opposite phenomenon obtains in languages with two or more non-synonymous words which we would translate indiscriminately as “death.” Italian, for example, has three words for death. Morte (= death, neutral) and decesso (= death at the end of illness) refer to essentially the same concept under different aspects, but trapasso (= passage from this life (on earth) to a better one (in paradise)) expresses a religious concept of death that the other words do not, namely that after bodily morte, personal life continues in a spiritual form. This is similar to the English euphemism “to pass on,” which implies passage to somewhere else, as opposed to “pass away,” which carries more overtones of disappearance from existence.
Tamil has three nouns for death. Irappu means departing from the world, shedding life (the opposite of pirappu, which means “birth,” “coming into the world with life”); it is a process. Chavu refers to the state of no life, and can refer to plants and animals as well as people. Maranam refers to the event of death and is used only in reference to humans [personal communication, A. S. Sundar, Etymologist,
The latter example is also a particularly clear instance of another way that death concepts and the whole death debate can be influenced by language: namely, the distinction that many languages, but not English, make between human and non-human death. Polish, for example, has two nouns for death. The most frequently used is śmierć (of common Slavonic origin), a non-marked equivalent of the English “death.” It is used in everyday speech as well as in legal and medical language, e.g. śmierć kliniczna (clinical death), śmierć mózgu (death of the brain). The other term, zgon, is not only more formal and medical, but means specifically “death of a person”: one may not say “zgon mózgu” (zgon/death of the brain) but only “śmierć mózgu” (śmierć/death of the brain); afterwards a doctor may declare zgon of the patient. Polish also has two separate words for “to die”: umrzeć, used only in reference to human beings, and zdechnąć, used in reference to animals or derogatorily in reference to humans. Similarly, the German verb D. A. SHEWMON AND E. S. SHEWMON 104
sterben applies only to humans (and personalized animals like pets); for animals one uses verenden, and for plants or animals, eingehen.
Hindi has three words for death: two from the same root, mrityu (Sanskrit) and maut (modern form of mrityu), and one from a different one, dehant (meaning literally “end of the body”). Although the three can be used relatively interchangeably, the interesting thing is that, in contrast to the aforementioned languages which have separate words for human vs. non-human death, these Hindi death-words apply equally to humans and animals, but not to plants. In English we can say that a plant “died,” but in Hindi they can only say that it “dried up” or “withered.” Thus the Hindi language both reflects and reinforces the Hindu belief in reincarnation among sentient forms of life.
The insistence by some, including the first author in his earlier thinking, following Bernat, that the correct concept of death must be species-nonspecific, applicable to all living things, might therefore derive more from a linguist accident (that English has only one proper word for death) than from anything else. The debate over whether human death should be conceived as primarily personal vs. biological might not even take place in languages that distinguish the two concepts of death by different words. (Whether “personal” can be reduced to “mental” is another whole issue.) English is surprisingly impoverished in this area, compared to its general richness of vocabulary and capacity for nuance of expression.
This view represents a radical departure from the first author’s previous under-standing of death as a species-nonspecific, unitary phenomenon of a fundamentally biological nature, to which other disciplines ought to refer. Defending “brain death” at the Pontifical Academy of Sciences in 1989, for example, he stated (in a subsection entitled “Death is a unitary phenomenon”): “The only alternative [to postulating multiple “kinds of death”], which is intuitively more reasonable anyway, is to regard human death as a singular reality, which can be considered from a variety of perspectives [biological, clinical, philosophical, theological, legal].”6(p.28) Even after his subsequent rejection of the traditional “organism as a whole” or “integrative unity” rationale for “brain death,”2,7,34, 41-43 he continued to subscribe to the “unitary phenomenon” assumption, up until the recent rethinking expounded here.
All that time he had assumed that the essence, in the biological realm, of the one phenomenon of death (cessation of life, regardless of form or species) could be expressed in at least two ways (two candidate “concepts of death,” in the tripartite terminology of Bernat et al.1): loss of endogenous opposition to entropy, and loss of integrative unity of the organism as a whole. (Maintaining the parallel with the empirical events labeled “E_”, let us call these conceptual moments C1 and C2, respectively.) It was taken as axiomatic that these two concepts were merely different ways of describing the same thing and therefore necessarily pointed to the same moment in time. Such “concepts” of death are not so abstract as others like “loss of the life principle,” which do not imply a physically definable moment along the biological time line. The single moment corresponding to these two physicalistic “concepts” was taken to be the “true” moment of death, which one should not assume a priori to correspond necessarily to any of the E’s listed in the previous section, some of which could at best be tolerated as token “moments of death” for the sake of cultural sensitivity or convenient legal fiction.
Now we must question whether either C1 or C2 is determinable or even meaningful on the empirical level; much less can their temporal coincidence be taken as axiomatic. SEMIOTICS OF DEATH 105
Moreover, there is no a priori reason to assume that either of these concepts is intrinsically “truer” or more “objective” than the death-concepts in cultures with no linguistic equivalent for the English word “death.”
4.1. C1 – Loss of Endogenous Opposition to Entropy
As noted in the beginning, “life” – like “unity,” “being,” and “consciousness” – is an “ur-phenomenon,” a fundamental concept which we come to know only through direct experience, not through more fundamental concepts to which it is reducible, of which there are none.3 Localized opposition to the general tendency in nature to increasing entropy is an extremely important aspect of physical life, but, as Seifert points out, it does not apply to all instances of life (e.g., spores, frozen goldfish), nor does it encompass the entire essence of life even in instances where it applies. If we accept a richer, undefinable but intuitable notion of life that cannot entirely be reduced to thermodynamics, we should expect a corresponding variety of types and manifestations of death, not entirely reducible to thermodynamics.
Let us take Seifert’s thesis one step further. It is simply a gratuitous assumption that entropy is a measurable – or even meaningful – property of living systems. It is a quantitative abstraction with different meanings in different contexts. Thermodynamic entropy is the amount of thermal energy unavailable to do work in a closed system. The Second Law of Thermodynamics states that entropy in a closed system can never decrease (i.e., available energy can never spontaneously arise out of nowhere). Logical or informational entropy is a measure of disorganization or disorder. A common-sense analog of the Second Law states that order does not spontaneously arise from disorder.
Without entering into the fascinating debate about entropy and the origin and evolution of life on earth, let it be simply noted that for any individual living organism there is no method or formula for calculating its entropy (whether thermodynamic or informational). It is not at all evident that such an abstraction, invented for the physics of gases, has any quantitative meaning in biology. Moreover, entropy (even if it could be measured) and its rate and direction of change are surely not homogeneously distributed within a dying organism; there are no doubt pockets or subsystems already breaking down and increasing in “disorder” while other parts continue to resist that tendency with greater or lesser success. There is simply no coherent way, even in a thought experiment, to quantitate the “entropy” of such a system. Moreover, in what sense is a terminally ill, irreversibly dying (but not yet “dead”) organism still “anti-entropic” anyway?
The entropy metaphor from physics sounds attractive and explanatory at first, but when its application to the transition from life to death is carefully examined, it is seen to be precisely a metaphor, not a more accurate description. There is no operationally meaningful way, even in principle, to determine a moment along the time course of dying/decaying when “endogenous opposition to entropy” instantaneously changes to “a tendency to increasing entropy.” C1 is nothing more than a mental construct, which for logical and linguistic reasons must refer to an imaginary moment in time and not a duration, but which lacks any clear physical correlate. D. A. SHEWMON AND E. S. SHEWMON 106
4.2. C2 – Cessation of the Organism as a Whole
In a recent article the first author proposed an operational definition for “integrative unity,” intended as a first pass which hopefully others would help to fine-tune.7 It was based on the implicit assumption, taken universally for granted in the “brain-death” debate, that unity (“organism as a whole”) vs. multiplicity (collection of organs and tissues) is a dichotomy translatable (at least in theory) from the philosophical to the physical domain. But if that assumption is incorrect, as we now suspect, then any attempt to operationally define “organism as a whole,” with the goal of enabling unequivocal, nonarbitrary, dichotomous categorization of all cases, is an exercise in futility.
To begin with, there are difficult questions regarding exactly what constitutes “the organism.” For example, what is the status (whether part of “the organism” or not) of an abscess, a cyst, a hamartoma, a benign tumor, an oocyte in an ovary or Fallopian tube, a conjoined twin with a common heart, the tip of a hair or fingernail, stones in the gall bladder, “normal” bacteria in the gut necessary for digestion, a transplanted organ, a molecule of ethanol in the bloodstream, a molecule of botulinum toxin in the bloodstream, antibodies causing automimmune disease, a partially amputated finger dangling from the hand by a thread of skin, etc. We are increasingly skeptical that there can be some all-encompassing, philosophically sound criterion for determining what, in medical practice, constitutes “the organism,” especially during its dying phase as various parts become dysfunctional and necrose at different times and rates.
That article cited a litany of properties “at the level of the whole,” some or many of which are present in at least some “brain-dead” bodies, from which it was concluded (and we would still conclude) that those bodies must therefore be “organisms as a whole.” The problem is how to apply this concept to bodies dying in the more usual cardio-pulmonary way, in which there is a relatively long window of time during the dying/decaying process when no holistic function is any longer evident but neither can it be demonstrated or inferred that all are absent (especially those that are not directly observable but exist as potencies). Take, for example, “capacity for wound healing,” an empirical test for which would take days, but the dying/decaying process might predictably reach the death-state with certainty within a few hours. How, then, can one determine at what moment the holistic property “capacity for wound healing,” along with many others like it, is lost?
The situation seems something like (though not on the same basis as) Gödelian unprovability. For a given set of axioms (viz., the death-concept axiom, in the logical domain, of intrinsic discontinuity between unity and multiplicity, between anti- vs. pro-“entropy”), there is an infinite number of statements that are neither provable nor disprovable (an infinite number of moments along the course of dying/decaying in a concrete case that can neither be proved nor disproved to correspond to the moment of discontinuity in the logical domain). Or it is like the fuzzy fractal border of certain mathematical structures such as the Mandelbrot set, for which there may be infinitely many points intrinsically impossible to determine whether inside or outside the set.44(pp.138-140)
A possible reaction to this unsatisfying lack of clear projection from a moment in the logical-conceptual domain to a moment in the physico-temporal domain might be to reject the philosophical notion of unity altogether – bringing us face to face with the SEMIOTICS OF DEATH 107
deepest controversies of ontology. We find it far preferable to say that healthy, living organisms are obviously integrated unities, that decomposing corpses are obviously not unities, and that there is a fuzzy area in between that is intrinsically undecidable. This is similar to how in color charts or prismatic spectra, two “adjacent” colors, e.g., green and blue, are clearly distinct and identifiable when their central areas are compared, but points along the transition zone are undecidable. The understanding of certain comparisons as dichotomous, when viewed on a large scale, such as between a healthy organism and a putrefying corpse, is not invalidated by the fuzziness of the transition viewed on a small scale, any more than the representational meaning of an impressionist painting is vitiated by the fact that close-up one sees only brush strokes.
4.3. Death as a Contextually Defined Event
Unlike E1-E7, which are clear, physically definable moments, C1 and C2 ultimately are abstract philosophical notions that do not translate or project operationally to the physical level. Although they may be good (possibly the best) candidates for an absolute, unitary concept of “death,” they are ironically no better than “departure of the soul” for leading practically to criteria and tests. Since neither the cessation of “anti-entropy” nor the cessation of the “organism as a whole” occurs at a determinable moment, a fortiori there is no reason to insist axiomatically that these two indeterminable moments should coincide. And (for those interested in the soul) what basis is there to suppose that the cessation of the soul’s “informing” the body corresponds to C1 or C2, if it seems intuitive that it ought to correspond to both, but the two may not necessarily correspond to each other? And if it might not correspond to one or the other, why must it necessarily correspond to either, and not to E1, E2, or …, depending on the context?
Perhaps a more mature approach to the question of the moment of death would be to recall Humboldt’s theory of the dynamic interaction between language and thought, and language and culture: perhaps our notion of death as a unitary phenomenon is after all not a self-evident “given” but an oversimplification reinforced by the fact that our language contains just a single word for death (namely, “death”). Much futile argumentation could be spared by recognizing that “death” is as rich and multi-layered a concept as “life” and equally undefinable. We should abandon the search for criteria for the universally “true” moment of death, as there is no single, context-independent, “true” moment of death. Rather, there are various moments of state-discontinuity, not all of which necessarily occur in a given case, and not all of which are equally striking to the senses and intellect of an observer. All of these state-discontinuities are equally “real” and “valid” phenomena in themselves, and there is no a priori reason that one of them must be singled out for the designation “death” while the others slip into conceptual obscurity for want of a word.
Once we recognize the restrictions that our language tends to impose on our ways of thinking about death, we can attempt to transcend them through expanding the vocabulary to correspond to the more enlightened understanding. We could invent words for E1, E2, etc., that were distinct enough not to create a false impression that they were all species of the same conceptual genus “death,” but simply different moments of state-discontinuity resulting from changes in observable parameters along the continuous process known as dying and decaying. D. A. SHEWMON AND E. S. SHEWMON 108
Depending on the clinical circumstances of the dying person and the behavioral or ethical issues for those standing round, one of these state-discontinuities will stand out as particularly observationally striking and/or ethically determining. Such an E could legitimately be labeled “death” for purposes of that clinical-ethical context. This approach is somewhat similar to proposals that we should “unbundle” what Veatch calls “death behaviors,” focusing not on when “death” occurs but on when such-and-such “death behavior” becomes appropriate or licit.45-48,49(p.26),50 We suggest that the candidates for death-moments be defined not purely by ethical propriety, but also by physical circumstances. Some moments might involve the same set of death behaviors but are legitimately distinguished from each other as equally valid “moments of death” or “death events” within their own contexts (e.g., E1 and E2).
5. IMPLICATIONS FOR TRANSPLANTATION
What implications does all this have for the “dead donor rule” and transplantation? Among the various types of “death behavior,” removal of vital organs is no doubt the one for which precision of timing is most critical. For non-transplant issues such as grieving, inheritance, burial, etc., it doesn’t really matter whether “death” is considered to occur at E1 or E2. If the timing is off even by hours, there are no practical consequences, except possibly in bizarre, highly improbable scenarios purposefully constructed to create a hypothetical consequence. Neither does the choice of E matter for burial, because by the time funeral services are arranged and burial actually takes place, it is typically a day or more after all of the Es. For ordinary purposes, it is reasonable to label as “death” (and to define as “legal death”) whichever E entails the most obvious state-discontinuity: in a home death that would probably be E1, whereas in a hospital death it could be E2.
Regarding transplantation of unpaired vital organs (or both of paired vital organs), the “dead donor rule” reflects the belief of many that the critical ethical question is: “Is the donor dead?” But if there is no one, true “moment of death” in an absolute sense, but rather a multiplicity of moments, any one of which might serve as a reasonable demarcation for a particular context, how does one decide which is most appropriate for the context of transplantation? We agree with Veatch and others that the proper choice of E is essentially not an ontological but a moral issue. “Is the patient dead?” is not only the wrong question to ask on the practical, physical level; it is not even a meaningful one when asked on a microscopic time-scale at the transition between life and death (like zooming in on the prismatic spectrum midway between green and blue, and demanding that someone not only identify it unequivocally as either “green” or “blue” but also have a convincing, logical rationale for doing so).
The question that really matters is: If we extirpate such-and-such organ(s) in such-and-such a way, do we kill or harm the patient? Although the verb “to kill” implicitly involves a dichotomous notion of life and death, it also involves causality and intentionality. The latter aspects make it possible in some situations to bypass the fuzzy, intrinsically undecidable border-zone between life and death, so that one can be morally certain of “not killing” even without first having to determine which side of the life-death boundary the donor is on at the time. SEMIOTICS OF DEATH 109
Truog, who agrees that a dead brain per se does not constitute a dead organism, maintains that removal of vital organs (especially the heart) from a heart-beating donor actively, iatrogenically kills the donor, but that it is nevertheless ethical because being killed supposedly does not “harm” the donor.51,52 He reaches this startling conclusion through a subjective, first-person-experiential (essentially Cartesian) notion of “harm,” according to which permanently unconscious patients are by definition beyond “harm.”
We do not subscribe to such a concept of either person or harm, and we believe that one can justify many transplantation actions without having to resort to such a redefinition, fully respecting the traditional injunction primum non nocere even in the sense of physical harm. It is beyond the present scope to enter into the philosophical debate over the fundamental principles of ethics. Let us therefore simply state without argument that we believe, contra Truog, that there is a profound and critical difference between killing and letting-die, as well as between intending and foreseeing death, even if in some pairs of examples the physical acts or omissions might look outwardly identical. We also believe that the principle of double effect is valid and necessary for bioethics, as applied eloquently in DuBois’s justification of certain controversial practices involving NHBD.53
DuBois zeroed in on the real issue in the very title of his article, “Is organ procurement causing the death of patients?”,53 in contrast to its counterpoint piece, “The importance of being dead: Non-heart-beating organ donation.”54 DuBois is right to suggest in his abstract “that many of the debates over death can be bypassed by changing the term of the debate: …” But what follows (“what matters is not whether death is a process or an event, but death as a state”) falls short of his own most fundamental point. It should rather be said that what matters is not whether death is a process or an event or a state, but whether “organ procurement caus[es] the death of patients,” as his title asks.
Let us restrict discussion to the ideal NHBD context: assuming the legitimacy of stopping life support (independent of transplant considerations), truly informed consent, lack of conflict of interest, medical certainty that apnea will supervene once the ventilator is discontinued, etc. When the ventilator is withdrawn in the operating room, the first E to occur will be E1, final apnea; then will ensue E2, E3, E4, and E6, in that order (presumably E5, the onset of loss of consciousness has already taken place, whether from primary brain damage or from sedation for the procedure). One should not base the timing of organ retrieval on which E represents “true” death, because this is an improperly posed question. Rather one should ask, “Beyond which of these events does the removal of organs X, Y, Z… neither kill nor harm the patient (even in the physical sense of accelerating the dying process)?” The answer depends on which organs one is talking about.
In the case of all organs except heart and lungs, removing them even before E1 will neither cause nor hasten death (however defined), because, in the ideal NHBD context under discussion (the ideality supposition must be repeatedly emphasized), by the time loss of those organs might exert even the tiniest systemic effect, all the E’s will have supervened long before anyway. It takes days or weeks to die from renal or hepatic failure or intestinal non-absorption or pancreatic insufficiency; for the first several hours, absence of such organs has no significant adverse effect on the body. Between discontinuation of the ventilator and even the most conservative choice of E, the effects of hypoxia-ischemia will totally overshadow whatever infinitesimal adverse effects might D. A. SHEWMON AND E. S. SHEWMON 110
theoretically result from an incipient lack of kidney or liver functioning, etc. Thus, for transplantation of non-cardiopulmonary organs, it is utterly irrelevant ethically whether “brain death” is “really death,” or whether the
For heart and/or lungs, the moral requirement “to do no harm” by the extirpation is trickier. A solution suggested by some would be to place the patient on cardiopulmonary bypass, remove heart and lungs without affecting systemic circulation or oxygenation, then declare the bypass machine an ethically inappropriate (“extraordinary” or “dispro-portionate”) means of life support (just as the original ventilator was), and disconnect it. Critics argue that this seems nothing more than ethical sleight of hand to pseudo-justify something that is morally equivalent to simply cutting out the beating heart. Even if such a roundabout, complicated and more expensive way of doing things really made a moral difference between unethical killing and ethical “letting die” (which we seriously doubt), such a procedure would still create the appearance of something shady. We doubt whether such an approach would do anything to relieve the moral stress and emotional burn-out experienced by operating room personnel involved in transplantation.55-57
An alternative approach that seems to respect traditional ethical principles and sensibilities could be the following. If, as we have been assuming for the ideal case, it is known for certain that (1) the patient will not breathe spontaneously off the ventilator and (2) it is moral to disconnect the ventilator, then it makes no physical or moral difference whether the ventilator is disconnected before or after opening the chest cavity. Therefore, go ahead and disconnect it with the chest cavity open, perhaps after already having removed some of the non-cardiopulmonary organs discussed above. Be prepared to instill directly into the heart and/or pulmonary circulation some cool, tissue-preserving fluid as soon as final asystole, E2, can be determined with moral certainty to be truly final. The organs could also have been pretreated with some tissue-preserving medication infused into the blood stream, as long as it wouldn’t diminish cardiopulmonary functioning and informed consent had been granted.
This approach to heart/lung retrieval does not cause or hasten death, because once circulation has effectively ceased due to the effect of progressive hypoxia on the heart, the dying or decaying process continues just the same regardless whether the nonbeating heart and nonfunctioning lungs remain physically in the circulationless body or not. The phrase “circulation has effectively ceased” is intentionally chosen, because what is important here, both physiologically and ethically, is the circulation of blood (even as its oxygen content progressively diminishes), not the QRS complex of the electrocardiogram, nor even a possible surprise “truly final” beat or two after (say) half a minute of asystole, beats which would not produce an “effective circulation” that would change in any significant way the process of dying/decaying already set in motion. Thus, it does not really matter exactly when E3 (loss of potential for cardiac auto-resuscitation) occurs or whether 2 or 5 minutes or some other duration of asystole suffices to provide certainty that E3 has already passed. This is because any potential cardiac auto-resuscitation would necessarily be very transient, by virtue of the ongoing apnea and severe hypoxemia; moreover, a very brief and weak circulation of anoxic blood would do nothing to counteract the inexorable process of dying/decaying already set in motion. In such a context, it makes no physiological difference in the dying process of the pulseless, SEMIOTICS OF DEATH 111
hypoxemic body whether the potential for cardiac auto-resuscitation is briefly actualized or not.
Therefore, as soon as (1) the blood becomes hypoxemic enough that any brief resumption of its circulation would make no physiological difference and (2) the cessation of effective heartbeat lasts long enough to make an experienced physician suspect that the last (in the sense of previous) heartbeat will prove to be the last (in the sense of final) physiologically effective one, then (a) there is no need to wait further for E3 in order to avoid causing or hastening death by cardiac removal, and (b) there is no moral problem in making the suspected E2 definitively E2 (via pre-E3 organ removal).
Transplant surgeons might complain that having to wait for E2 would only complicate their already difficult task and might risk ischemic damage to the donated organ(s). Perhaps the old adage “necessity is the mother of invention” applies here: transplant surgeons never developed such a technique for heart-lung retrieval, primarily because the “brain-death” fiction convinced them that there was no need to. We believe that a historically honest and physiologically enlightened appraisal of “brain death” makes it an ethical requisite. With a little creativity, the complication of waiting for E2 can surely be effectively dealt with – and for the sake of everyone’s consciences, it is a complication worth accepting and dealing with.
Statutory law does not require donors to be dead, but it does contain strong injunctions against voluntary homicide. In the transplant setting, these laws can be respected without needing to gerrymander a special statutory definition of “death” in order to seemingly fulfill a “dead donor rule” that isn’t even a law. Statutory law would benefit by refocusing from the wrong medical-ethical question (“When is the patient dead?”) to the right one, which it already has ample provision for (“Is homicide being committed?”).
5.1. Disclaimer
To avoid potential misinterpretation, let it be stated explicitly that the foregoing discussion does not constitute positive advocacy of any particular transplantation protocol, whether heart-beating or non-heart-beating, present or future. The above analysis addresses the very precise and limited question whether it is possible in principle to remove vital organs without causing or hastening death or violating the time-honored injunction primum non nocere. Our conclusion is: yes, it is possible in principle. But before deciding whether it would be prudent to put this principle into practice in today’s society, many other factors must be considered which are outside the scope of this discussion – such as whether donor consent can be guaranteed to be truly informed and free, whether in the case at hand apnea off life-support can be predicted with medical and moral certainty, whether such eviscerating procedures respect human dignity even if they might not cause or hasten death, whether the risk of public misperception that this is utilitarian killing can be minimized, etc. If the answer to one or more of these “whethers” is “no,” as critics of NHBD protocols claim,58 then it behooves us to hold off implementing the otherwise intrinsically ethical procedure until all the circumstantial details are worked out. The foregoing section is merely an appeal for clarity of terminology and logicality of thought, absolute prerequisites for fruitful discussion and valid ethics. D. A. SHEWMON AND E. S. SHEWMON 112
6. CONCLUSION
Linguistic considerations surrounding the term and concept “death” suggest that society has traditionally assumed a univocal notion of “death,” in large part because, up until very recently in human history, there was no need for a more nuanced notion. Thus, the English language developed with only a single word for death (namely, “death” and its relatives “dead,” “to die,” etc. – euphemisms, contextual and stylistic variants excluded). What served humankind well linguistically for most of history now tends to restrict thinking when applied to situations uniquely occasioned by modern medicine.
It is not a contradiction in terms to speak of state-discontinuities brought about by continuous changes in observable parameters; bifurcations in state-space are ubiquitous throughout nature, especially animate nature. The reality of death-processes therefore does not preclude the reality of death-events. There is no logical continuum between life and non-life, being and non-being, unity and multiplicity, and it is simply incorrect even to pose the tiresome question whether “death” is an event or a process. Linguistically it can be understood only as an event; there are other words for the process. It is time to expand our death-vocabulary to facilitate the recognition of multiple events, all equally real, along the process from declining health to decomposition.
Depending on the context, some of these death-related events may constitute a more obvious discontinuity than others and may more justifiably be considered “death” within that context. It may also be more appropriate emotionally and/or morally to begin certain kinds of “death behavior” at one of these moments and not others, depending on the clinical context and the behavior in question. There is no reason to assume a priori that there must be an overarching, unitary conception of death from which all diagnostic criteria and tests must derive.
Regarding organ transplantation, the important and truly meaningful question is not “When is the patient dead?” but rather “When can organs X, Y, Z… be removed without causing or hastening death or harming the patient in any way?” Perhaps some of the general public’s confusion and incoherence surrounding the “dead donor rule” results from a mismatch between people’s intuitive understanding of death in the era of modern medicine and the limited lexicon that our colloquial language imposes on us for articulating that intuitive understanding.
7. ACKNOWLEDGMENTS
The first author thanks Alan Garfinkel, Ph.D. for introducing him to nonlinear dynamics and the example of the Hopf bifurcation; and Stuart Youngner, M.D. and Laura Siminoff, Ph.D. for catalyzing this latest stage in his understanding of death. Both authors thank the many colleagues from the “Linguist List” (http://www.linguistlist.org/) who replied to our query about “death words” in other languages. SEMIOTICS OF DEATH 113
8. REFERENCES
1. Bernat JL, Culver CM, Gert B. On the definition and criterion of death. Ann Intern Med 1981;94(3):389-394.
2. Shewmon DA. "Brain-stem death", "brain death" and death: a critical re-evaluation of the purported evidence. Issues Law Med 1998;14(2):125-145.
3. Seifert J. What is Life? The Originality, Irreducibility, and Value of Life.
4. Schrödinger E. What is Life? The Physical Aspect of the Living Cell.
5. Korein J. The problem of brain death: development and history. Ann N Y Acad Sci 1978;315:19-38.
6. Shewmon DA. "Brain death": a valid theme with invalid variations, blurred by semantic ambiguity. In: White RJ, Angstwurm H, Carrasco de Paula I, eds. Working Group on the Determination of Brain Death and its Relationship to Human Death. 10-14 December, 1989. (Scripta Varia 83).
7. Shewmon DA. The brain and somatic integration: insights into the standard biological rationale for equating "brain death" with death. J Med Philos 2001;26(5):457-478.
8.
9. Morison RS. Death: Process or event? Science 1971;173:694-698.
10. Kass LR. Death as an event: a commentary on Robert Morison. Attempts to blur the distinction between a man alive and a man dead are both unsound and dangerous. Science 1971;173:698-702.
11. Emanuel LL. Reexamining death: the asymptotic model and a bounded zone definition.
12. Varela FJ. Principles of Biological Autonomy.
13. Hilborn RC. Chaos and Nonlinear Dynamics. An Introduction for Scientists and Engineers.
14. Aihara K, Matsumoto G. Chaotic oscillations and bifurcations in squid giant axons. In: Holden AV, ed. Chaos. Princeton:
15. Garfinkel A. A mathematics for physiology. Am J Physiol 1983;245(Regulatory Integrative Comp Physiol 14):R455-R466.
16. Garfinkel A, Chen P-S, Walter DO, Karagueuzian HS, Kogan B, Evans SJ, Karpoukhin M, Hwang C, Uchida T, Gotoh M, Nwasokwa O, Sager P, Weiss JN. Quasiperiodicity and chaos in cardiac fibrillation. J Clin Invest 1997;99(2):305-314.
17. Glass L, Guevara MR, Shrier A, Perez R. Bifurcation and chaos in a periodically stimulated cardiac oscillator. Physica 1983;7D:89-101.
18. Strogatz SH. Nonlinear Dynamics and Chaos, with Applications to Physics, Biology, Chemistry, and Engineering.
19. Botne R. To die across languages: Toward a typology of achievement verbs. Linguistic Typology 2003;7(2): 233–278.
20. DuBois. Non-heart-beating organ donation: A defense of the required determination of death. J Law Med Ethics 1999;27:126-136.
21. Fortescue M. West Geenlandic.
22. Lynn J,
23. Potts M, Byrne PA, Nilges RG. Beyond Brain Death. The Case Against Brain Based Criteria for Human Death.
24. Singer P. Rethinking Life & Death. The Collapse of Our Traditional Ethics.
25. Youngner SJ, Arnold RM, Schapiro R. The Definition of Death: Contemporary Controversies.
26. Rosner F. Modern Medicine and Jewish Ethics.
27. Rosner F. The definition of death in Jewish law. In: Youngner SJ, Arnold RM, Schapiro R, eds. The Definition of Death: Contemporary Controversies.
28. Steinberg A. Ethical issues in nephrology - Jewish perspectives. Nephrol Dial Transplant 1996;11:961-963.
29. Yofe YA, Mark Y. Groyser verterbukh fun der yidisher shprakh.
30. Cole DJ. The reversibility of death. J Med Ethics 1992;18(1):26-30; discussion 31-33.
31. Lynn J. Are the patients who become organ donors under the
32. Tomlinson T. The irreversibility of death: reply to Cole. In:
33. Cole D. Statutory definitions of death and the management of terminally ill patients who may become donors. In:
34. Shewmon DA. Spinal shock and 'brain death': somatic pathophysiological equivalence and implications for the integrative-unity rationale. Spinal Cord 1999;37(5):313-324.
35. Bernat JL. How much of the brain must die in brain death? J Clin Ethics 1992;3(1):21-26; discussion 27-28.
36. Bernat JL. A defense of the whole-brain concept of death.
37. Pritchard ET. No Word for Time: The Way of the Algonquin People.
38. Cook JA. The Hebrew verb: a grammaticalization approach. Zeitschrift für Althebraistik (
39.
40. Elliot E. Through Gates of Splendor.
41. Shewmon DA. Recovery from "brain death": A neurologist's Apologia. Linacre Q 1997; 64(1):30-96.
42. Shewmon DA. Chronic “brain death”: meta-analysis and conceptual consequences. Neurology 1998;51(6): 1538-1545.
43. Shewmon DA. Chronic "brain death": meta-analysis and conceptual consequences [response to letters]. Neurology 1999;53(6):1369-1372.
44. Penrose R. The Emperor's New Mind.
45. Brody BA. How much of the brain must be dead? In: Youngner SJ, Arnold RM, Schapiro R, eds. The Definition of Death: Contemporary Controversies.
46. Fost N. The unimportance of death. In: Youngner SJ, Arnold RM, Schapiro R, eds. The Definition of Death: Contemporary Controversies.
47. Halevy A. Beyond brain death? J Med Philos 2001;26(5):493-501.
48. Halevy A, Brody B. Brain death: reconciling definitions, criteria, and tests. Ann Intern Med 1993;119(6): 519-525.
49.
50. Youngner SJ, Arnold RM. Philosophical debates about the definition of death: Who cares? J Med Philos 2001;26(5):527-537.
52. Truog RD, Robinson WM. Role of brain death and the dead-donor rule in the ethics of organ transplantation. Crit Care Med 2003;31(9):2391-2396.
53. DuBois JM. Is organ procurement causing the death of patients? Issues Law Med 2002;18(1):21-41.
54. Menikoff J. The importance of being dead: Non-heart-beating organ donation. Issues Law Med 2002;18(1): 3-20.
55. Castelnuovo-Tedesco P. Cardiac surgeons look at transplantation - interviews with Drs. Cleveland, Cooley, DeBakey, Hallman and Rochelle. Semin Psychiatry 1971;3(1):5-16.
56. Lock M. Twice Dead: Organ Transplants and the Reinvention of Death.
57. Youngner SJ, Allen M, Bartlett ET, Cascorbi HF, Hau T,
58. Arnold RM, Youngner SJ, Schapiro R, Spicer CM. Procuring organs for transplant: the debate over non-heart-beating cadaver protocols.
No comments:
Post a Comment