a) General prevention

As far as general prevention (positive or negative) is concerned, life imprisonment can be justified only if it is empirically demonstrated that it prevents crimes (^{Cuneo, 2017}, p. 100). This line of thinking comes up against two practical issues: on the one hand, such compatibility seems difficult because experience shows that crimes punishable for long periods of time with severe penalties continue to be committed, making it problematic to reconcile any penalty with general prevention. And, then again, it is difficult to define whether it is the penalty that generates a preventive effect, or whether, most people prefer not to commit a crime not out of any dread of punishment, but out of other ethical considerations. While general prevention does not appear to encounter axiological impediments to justifying and defending life imprisonment, it does seem difficult to determine that this sanction does indeed lead to the anticipated outcome.

Already with Beccaria it was said that crime prevention is achieved more by the certainty of the penalty than by its severity. Therefore, if the Lombard Marquis is correct, the question of giving legitimacy to life imprisonment would appear to be complex, at least from the perspective of general prevention.³ On the same line of thought, according to Silvia Sanchez, "while increases in the severity of penalties have no proven preventive effects, those are visible when reinforcing the factors influencing greater certainty of the sanction (...) if the increase of certainty of criminal sanctions (and therefore, in general, the increase of certainty that the penal system will function satisfactorily), means verifiable increases in the preventive effectiveness of the rules, it indicates that criminal law as a whole intimidates more than its absence would intimídate» (^{Silva Sánchez, 2010}, pp. 252-253).⁴

The key issue of the empirical justification of life imprisonment lies in the fact that it has only been applied to crimes of high severity and it is precisely the perpetrators of these crimes who are usually difficult to intimidate; therefore, the intimidating threat will hardly produce the desired effects of this penalty. Paradoxically, it would be possible to conclude that life imprisonment would be justifiable, from a general prevention point of view, if drunken driving or a white-collar crime were punished with this penalty. On the other hand, it would appear difficult to justify it with respect to rape and murder.

Nor do we dare to go as far as to claim that this penalty would ensure the social peace sought after or desired by positive general prevention would indeed be achieved, although it could be argued, on citing specific earlier cases, that this does occur. In this respect, we do not know of any empirical studies that could prove this positive upshot of general positive prevention and we cannot imagine how they could be elaborated. To conclude, we believe that life imprisonment cannot be justified on the basis of negative general preventive criteria and we are unable to offer a categorical answer on its positive modality.

b) Special prevention

In principle, special negative prevention could justify life imprisonment, because the subject locked up in prison will not be able to perpetrate crimes against the free community (^{Cuneo, 2017}, p. 104). However, many crimes are committed in prison, from robberies and thefts, to forgeries, illegal traffic of alcohol and drugs, rapes and homicides. Today, in addition, multiple crimes are committed from inside the prisons towards the free community; for example, telephone scams, currency forgery, drug trafficking, etc. Moreover, organised crime groups, mafia and drug traffickers, usually recruit their soldiers in prison, who, once free, due to lack of opportunities, rapidly enter the frontlines of these organised crime groups. In those circumstances, it appears chimerical to affirm that the privation of liberty means that the inmate will not commit crimes. Prisons are a long way off from being a place where only a custodial sentence is served in which their prisoners live together without committing crimes, preparing for longed-for social reinsertion. Rather, on the contrary: levels of violence and the constant occurrence of crimes in the majority of prisons are far higher than in the free community. However, it is clear that, at least with respect to free society, the possibilities of committing crimes from prison are much lower than those of a free individual. Therefore, if this theory is capable of denying the dignity of those persons deprived of liberty, life imprisonment may well be justified for them.

According to special positive prevention, life imprisonment is justified only if we consider the existence of individuals whose re-socialisation is impossible, because, by separating the offender for life, it does not give him/her the opportunity to re-socialize. That is to say, this theory would justify life imprisonment only with respect to irredeemable criminals. That is to say in order for this theory, professing resocialisation, to justify life punishment, it is necessary for it to creates the existence of inferior human beings that need to be separated from social coexistence.

2.2 Retribution

According to retribution, life imprisonment cannot be justified, and this is on account of two interrelated reasons. First of all, because if retribution, being as it is degrading, regards criminals as people responsible for their actions, it should still see them as such when condemning them and throughout the time they serve their sentence. That is to say, according to retribution, criminals are indeed people as they respond for their own actions and when responding for them, that is to say, when they are serving the sentence they continue to be people, therefore penalties dehumanising convicts are omitted. The second reason as to why we maintain that retribution cannot justify life imprisonment is because retribution at the time of the sanction targets the I, the subject's identity, which, over the course of time changes, meaning the individual may no longer be punished for a crime committed long ago. We will start by analysing the first reason, while the second issue, concerning the individual's change of identity over the course of time, is to be discussed towards the end or our argumentation.

a) Retribution necessarily leads to a notion of an anthropologically-founded criminal law, a criminal law of people and for the people⁵, in which one cannot fail to see oneself in the convict as a fellow being. For this reason, life punishments are inadmissible, owing to their radical inhumanity, in that the concept of retribution itself envisages the person as an end, not as a means. This idea contrasts with both those who propose, justify or legitimise approaches that see the existence of human beings who are not persons as well as with a so-called criminal law of the enemy.⁶

This first limitation is based on a principle of humanity that concurs with every understood retribution. On the basis of security and through it, one comes to humanity, a community in its purest form, to cohabitation, which is not merely restricted to coexistence. The principle of humanity also implicates a sense of solidarity, individualism, which is not selfish, and leads to the realization of humanity both collectively and individually-speaking. It is not a humanity understood as one of fine sentiments, sentimentalism, tender-heartedness nor one speaking in weepy terms, but a humanity of Criminal Law drawn up objectively. This humanity, although its meaning is collective, culminates in an individual sense, because it conceives the human being as a person with its own individual dignity, holder of a personal and non-transferable destiny, with a capacity to dream and carry out their own actions with a series of implicit possibilities making that community, that coexistence, possible.

It is precisely this principle of humanity that makes Criminal Law meaningful. As such, Criminal Law according to this liberal conception, should be minimal, of ultima ratio, and should lead to a reductive, minimal meaning: when punishing it should apply the minimal affliction, only that which is indispensable. A liberal criminal law, humanised, that is applied to a community of free human beings, must necessarily be a reduced and fragmentary criminal law. An authoritarian criminal law, on the contrary, is not fragmentary. The least that can be done, by virtue of the principle of humanity, is to reduce penalties, decriminalise, depenalise, dejudicialise behaviour. We need to extract things from the Penal Law and foster models of conflict resolution. By virtue of the principle of humanity, punishments that treat criminals as non-people that exclude them, such as the death penalty or imprisonment, which in more or less prolonged periods breaks with cohabitation and nullifies personality, should be ruled out. A prolonged imprisonment, beyond certain limits - e.g. 15 years - causes severe and irreversible harm to the personality. Sentences of twenty or more years are, in fact, perpetual, because, although one returns to freedom, one becomes annihilated.

Although no one refutes the principle of humanity outright, this is deformed by the different ways in which this is adopted by the various legal systems are adopted. Preventivist doctrines disagree with this principle, given they degrade humanity and the dignity of the human being, transforming it into a mere instrument for goals set by others. With retribution, on the contrary, the devaluation of punishment is however achieved without the convict being excluded from cohabitation.

Goethe's words should here be taken into account, in the sense that in a community where human beings are free and as ends in themselves, whether it is to punish or to be soft, people should be treated with humanity. And being an end in itself means being a person with dignity. Acknowledging dignity forms the basis for equality, assuming that with dignity, no one is more dignified than the next person, but we are all equally dignified. With dignity hierarchies are excluded. A criminal law that is not liberal is only a phenomenon of power, it is pure punitive power and not law.

With regards to the maximum limits that a custodial sentence must entail (necessary for a proper retribution), these cannot be such as to cause irretrievable physical and psychological degradation. If we conclude that a sentence of more than 15 years of privation of liberty has a dehumanising impact on the criminal, then the penalty cannot exceed aforesaid time.

According to Von Hirsch, who puts forward a theory of deserving applied in a benign form, prison should be limited to serious crimes (mainly violent crimes and the most serious white-collars crimes), and the duration of the limiting of freedom for those crime should be no more than 3 years- except for murder, where 5 years would be the normal limit- (^{Von Hirsch, 1998}, p. 80). These limits are based, according to Von Hirsch, on psychological studies that account for the deterioration generated by the prison sentence (^{Von Hirsch, 1998}). On the other hand, Kleinig, based not so much on a theory of deserving as on the idea of the humane and decent treatment of human beings, proposes as a maximum penalty of 25 years (^{Von Hirsch, 1998}, p. 74).

b) Retribution targets the I, at the personal identity of the subject, which changes with time. Given what has previously been discussed, it seems pertinent to ask ourselves a further question on the personal identity of the I.

Fortunately, criminal law nowadays does not usually discuss the validity of a principle called the personality of penalties, according to which the sentence must be imposed on the subject considered responsible for the crime, who is the one that has to bear it (^{Silva Sánchez, 2008}, p. 662).⁷ This principle, together with the one on responsibility for one's own action, integrates the principle of guilt in a broad sense (^{Silva Sánchez, 2008}, p. 662) according to which only the same individual may be punished that committed the crime and who has been found responsible after a trial. On the contrary, when the penalty falls onto a different subject than the one who committed the crime, the principle of culpability is infringed. The requirement of identity of the subject compares individual identity to physical identity. Apparently, what lies behind the requirement of personal identity and the principle of guilt is the idea of deserving, essential to any retribution.⁸ But what is personal identity?

To answer this question, at least two philosophical schools of thought should be analysed. On one side, the empiricists or reductionists, who based on the physical and psychic changes that the person undergoes over the course of time, deny the subject's personal identity in different moments.⁹ According to them, not being able to assert the existence of a personal identity, at the most we could speak of a «mere community of successive mental states, phases or personal stages" (^{Silva Sánchez, 2008}, p. 666). In line with this thinking it would be appropriate to consider whether it is possible to hold criminally liable a subject in a time subsequent to that when the offence was committed (in the 2nd moment), seeing them in terms of the successor of that apparently same subject for a crime committed at a previous time (in the 1st moment). Since it seems difficult to argue that they should not be held accountable, the most pertinent question, in our opinion, would be whether or not we could reduce or annul this imputation on the basis of some change (time lapse or a traumatic event).

According to the other line of thinking, metaphysics, personal identity does indeed exist in the subject at different times. There is "a substantial personal identity based on the permanent presence of corporal and spiritual components in the subject (...) (^{Silva Sánchez, 2008}, p. 667). it is an essential permanent presence. As will be understood, according to this thesis neither time nor transcendent events would alter the individual essence. Therefore, there would be no problem in attributing responsibility for a past crime, even if several decades have passed, since it would be the same identical subject. Usually, for different legal-penal systems «personal identity is imputed» (^{Silva Sánchez, 2008}, p. 670). However, for practical reasons and those of justice, it seems necessary to limit such a normativist concept. Criminal doctrine overlooks the fact that the subject does change and assumes that the subject is always the same (so it is imputed); the time limits that are established in criminal law are related to practical reasons rather than to the supposed change of the I.

According to Merkel, the tacit demands of criminal prosecutions imply the assumption that, over the course of time, the subjects of significant criminal proceedings remain identical to themselves as persons (^{Merkel, 1999}, p. 502). However, this author is of the opinion that, in certain situations, this statement is false or implausible. To address the problem of identity in time - which, according to Merkel, has not been sufficiently developed by criminal science - we should recollect the Greek myth, told by Plutarch, of Theseus's ship, who on his return from Crete had the vessel's broken boards removed one by one, replacing them with new ones, to the point where he no longer preserved any of the originals. From this, the question arises whether it would still be the same ship after all the boards have been replaced. Merkel then adds a variant to the example: a sailor then collects all the parts that have been discarded to be replaced by new ones and builds a ship exactly the same as Theseus's. Which one would then be the original ship? For Merkel there could be three possible answers: the first one, the second or neither, since logically they could not be both at the same time. At the heart of the question lies the dilemma of whether, once the parts that make up a whole have changed, the identity is still there. If we transfer this example to human beings, it may be argued that, at least on a physical-biological level, people do not maintain a single cell throughout their lives, as they are all constantly undergoing renewal. Are we still then the same person after a certain period of time or after a certain amount of changes?¹⁰

Furthermore, what about the changes we experience at a psychological level? Do they also change our identify of the I? On this subject, Merkel analyses cases of Alzheimer and total loss of memory. Is an individual still the same person after complete loss of autonomy or any connection with his/her past? And, linking this to our own investigation, can we sentence someone who has no memory of their crime? For us, this question is also important with regards to when the sentence is implemented. If the offender loses their memory, can they continue to serve the sentence for something that is not recalled? Can we affirm that we continue inflict retribution on someone who does not know why he/she is serving a sentence?¹¹

If we are to understand that with time just deserts, the verdict meting out blame and the need for punishment diminish, then we should be asking ourselves when or in what circumstances the culpability of the subject in the past vanishes to such an extent that the punishment does not seem to be fair.¹² Both metaphysical and empirical theories, as they work on totally abstract levels, do not appear to satisfy the legal-criminal issues of the real world, such as the affliction of the prisoner suffering from loneliness in their cell. However, we may ascribe ourselves to an eclectic theory, understanding that there would be no problem to impute to a subject his previous actions with certain temporal or circumstantial limits;¹³ that is to say, we understand, from retribution, that the subject must answer for their past actions, but not indefinitely and perpetually.

The point is to think about guilt in a dynamic way, bearing in mind that the passing of time ensures the culpability of the subject is not same culpability attributed to him at the moment of sentencing and for which he/she was convicted in order to compensate for the deed that constituted a crime. This is a structural limitation that is understood to be meaningless if someone is sanctioned for the rest of his/her life or for excessively long periods.

Even if we are aware that the individual changes with time, we are not sure whether it is possible to say that at any moment this person is no longer the same who committed the crime. Something leads us to think that, after 20 years of the passing of sentence, he/she is no longer the same person who committed the crime. Clearly neither North American legislation nor several U.S. courts concur with our thesis, which repeatedly sentence minors between the ages of thirteen and seventeen to life imprisonment without the possibility of parole.

A study undertaken by Amnesty International and Human Rights Watch, published in 2005, shows that there are, in the United States, at last 2,225 people serving life imprisonment without parole for crimes committed when under the age of 18, for most of whom it was their first offense. 16% of those convicted would have committed their crime when aged between thirteen and fifteen (^{Human Right Watch, 2005}). The dramatic situation outlined in Amnesty International's and Human Rights Watch's report seems difficult to defend from a preventive point of view, although in relation to what we are dealing with now, this situation is unacceptable for a retributionist standpoint and not only because of the evident inhumanity of these punishments, since if the sentence is based on retribution, we must understand we are continuing to inflict punishment for the crime throughout the entire serving of the prison sentence. This forms the basis for our idea that retribution would then, at some point, be an undeserved because culpability, harnessed to personal identity, has diminished over time and with the change of identity of the subject, reaching a time to be so minimal as to no longer be able to justify the penalty.

After twenty or forty years an individual's core identity may still be there, but this core identity does not appear sufficient to justify the logic of responsibility, the logic of deservedness, the logic of retribution, because there would not be a sufficiently strong identity linking the same person who committed the crime with that serving the sentence at a much later stage in time. If we affirm that the subject changes, as his personal identity keeps on changing, there will come a time when it is no longer possible to continue compensating culpability because the person has come to have a different identity.

Our stated idea here professes that this subject cannot be held responsible after a long period of time has elapsed, because they would no longer be the same. Life imprisonment or a long sentence would involve an excess of punishment that could not be compensated by culpability, because culpability, understood as that part of oneself, determined by one's own will, by one's I that at a given moment has decided to commit an act of delinquency, is no longer the same. Although our theory appears to be a mere intuition, as it is impossible for us to empirically know if and when a person happens to have a different identity, we may attempt to make our assumptions of a scientific nature by linking them to studies on the phases of biological and psychological development.¹⁴

Erik Erikson, a German psychoanalyst (1902-1994), created in the United States a personality development theory in which he describes eight stages of the psychosocial development of the human being (^{Erikson, 1980}, p. 80). According to Erikson, a psychologically sane person should go through each one of those phases and suffer what is characteristic of each stage as well as enjoy its benefits.

The first four stages relate to crises and conflicts in the development of life up to the age of thirteen. These are: 1. - Trust versus Mistrust, a stage that occurs from birth to eighteen months of age approximately; 2. - Autonomy versus Shame and Doubt, a stage that occurs from eighteen months to three years; 3 - Initiative versus Guilt, a stage that occurs from three to five years old; and 4. Industry versus Inferiority, a stage that occurs from five years to thirteen years old approximately.

The last four phases are those that matter the most to us in order to understand a limitation of liability of a legal-criminal nature, considering that before the age of thirteen it makes no sense to attribute any type of criminal liability. This could respond to the assertion that until adolescence, the individual cannot develop the requirements of physiological development, mental maturation and social responsibility to experience and go through the identity crisis (^{Erikson, 1980}, p. 78). Let us look at the phases that-include life from adolescence to old age.

The fifth phase or Adolescence is characterised by the existence of a conflict of Identify Versus Confusion of Identity. It would cover the period the goes between thirteen and twenty years old. During this phase the young person seeks to define their own identity, it is a way of life between childhood and adulthood. Thus, in the last years of schooling, young people, harassed by the physiological revolution of their genital maturation and the uncertainty of their adult roles, seem very busy with capricious attempts to establish an adolescent subculture through what looks like a final identity (^{Erikson, 1980}, p. 110)^. The sixth phase, the young adult, lives a conflict of intimacy versus isolation. This would cover the period that goes between twenty-one years up to forty years of age approximately and it is characterised by greater confidence, as it seems that this young adult would no longer need, as before, to prove something to themselves (^{Erikson, 1980}, p. 111). The seventh phase, of adulthood, is characterised by a conflict of generativity versus stagnation, and would stretch from the forties to the sixties. This stage is usually related to parenting and is characterised by a constant concern for the future (of children, of future generations, etc.).

Lastly, in the eighth stage, of senescence or maturity, a conflict of integrity versus despair. This phase would cover the period from the age of sixty onwards, and is characterised by physical deterioration, the presence of disease, muscle weakening and a certain concern for one's own death or the death of peers. If philosophically we can justify the idea that the subject changes with time, that this affects culpability and furthermore that psychosocially these phases exist and have been analysed and studied by psychology, then it seems that what was only our intuition has a scientific basis enabling us to limit the sentence given as retribution, limit that would be around, according to the studies of Erikson, 20 years or less, that is to say, the duration of each stage (^{Erikson, 1980}).¹⁵ This arises from the need for the individual to be able to live in freedom at each stage in order to aspire to full development.

From the retribution perspective it seems coherent that the evil of punishment is prolonged only in one of the four stages: that the crime committed at one stage of life be punished at the same stage as when the crime was committed. Exceptionally, if the crime was committed at the end of one of the phases, at nineteen for example, the sentence could pass to the next stage taking its age-frame as its point of reference. However, it would appear prudent to limit the sentence and in no case to exceed half of the next stage, thus leaving half of this cycle free for the subject to develop during this new phase. In this way, the principle of personality of the penalties will be respected, avoiding that the punishment is applied to a person who may have undergone radical changes in his or her personality.