Introduction

The use of criminal law, as the most coercive means available to the state (or other entities bearing ius puniendi) to shape individual behaviour, must be accompanied by one question: what is its purpose? What function should criminal law serve? In short, what legitimises the state’s restriction of the most fundamental rights of the individuals who make up its social base?

The answer to this question is neither simple nor unambiguous. And although some parallelism exists between the questions ‘Why punish?’ and ‘What to punish?’, the two questions are distinct. In this short post, I will focus mainly on the second question: what should be criminalised? What lends legitimacy to this state intervention? Ultimately, how can the existence of punishment and its application to individuals be justified?

Several answers have been given to this question, some of which will be explored below. However, before doing so, it is necessary to introduce the concept of legitimacy itself, given its relevance to criminal law. Legitimacy can be understood through two conceptual binomials which express the different aspects of legitimate (criminal) law: input and output legitimacy; formal and material legitimacy[i].

The former concerns the assessment of the rule (and its adoption process) at two points in time. In order to be legitimate, criminal law must allow citizens to participate in the process of its formation – that is to say, it must be adopted democratically, which implies the decisive participation of Parliament, the representative body of the people. At the other end of the spectrum is the assessment of the norm by comparing the expected results with the results actually obtained (namely, by setting out the reasons justifying the adoption of the criminal norm and subsequently assessing its effects). Neither of these aspects is particularly linked to the content of the criminal norm.

The second vector relates to the evaluation of the norm in terms of compliance with the legally established procedure for its adoption (formal legitimacy) and in terms of its content (material legitimacy). Therefore, if the due legislative process is not followed, the criminal norm will be illegitimate; likewise, if the criminalisation of conduct is not directed at something legitimate, it should not be adopted (or, if adopted, it should be removed from the legal system). Taken together, these aspects contribute to the formation of the subjective conviction that a rule or institution is due obedience[ii].

The next sections will focus on the dimension of material legitimacy, addressing three distinct perspectives: the principle of the exclusive protection of legal goods, the harm principle, and legal moralism. This will ultimately enable an informed decision to be made regarding the theoretical basis that should underlie the choice to criminalise (or not) a particular behaviour, depending on whether that criminalisation is directed at something legitimate or illegitimate according to the adopted principle.

The principle of the exclusive protection of legal goods

The concept of the ‘legal good’ (in the original, Rechtsgut) as a limit (or rather, a prerequisite) to the legitimacy of criminal legislative intervention emerged in Germany with Birnbaum[iii], as a reaction to Feuerbach’s proposal to consider legitimate only those criminal rules that offended a subjective right[iv]. Birnbaum argued that crime did not truly affect a particular right, but rather the object of that right, and therefore this could not be the determining criterion for the legitimacy of criminal law (for example, in a crime of theft, it is not the right to property that is affected by criminal conduct – that right remains intact – but rather the property itself).

The evolution of the concept of legal good is beyond the scope of this discussion, but this initial idea led, due to the subsequent historical context, to the consideration of two possibilities. On the one hand, legal good would be anything and everything that the legislator deemed worthy of criminal protection (a view which was, in particular, championed by Binding[v], from a markedly positivist perspective). On the other hand, from a more naturalistic perspective, legal interests would be defined as only those interests deserving of legal protection – this choice, made by the legislator, would effectively transform a real-life situation into a legal one (a position put forward by von Liszt[vi]). Although anchored in different theoretical and philosophical foundations, the evolution of the concept has always oscillated within the spectrum formed by these two opposing premises: either the criminal legislator was  considered free to choose the conduct to be criminalised, since the fundamental task of defining the interests deserving legal protection (and, by extension, criminal law protection) could not be taken away from them; or it was recognised that this freedom of the legislator should have certain limits – with authors diverging on the source of these limits –, outside of which the choice of interests to be protected would render the criminal law illegitimate.

Within the framework of the contemporary Portuguese criminal justice system (which will be used for the purposes of providing an example[vii]) imposing limits on the state’s ius puniendi is an unavoidable condition for the legitimacy of criminal law. However, for these limits to be effective, they must not originate from the criminal justice system itself. Otherwise, the critical function of legal goods (in the sense of assessing whether the content of a given rule is legitimate or not) would not be fulfilled, as the criminal justice system would be defining the limits of its own legitimate action. These limits must therefore be found outside that system.

Several suggestions have in the meantime emerged, which would lead to a more or less broad criminal law (and with more or less possibilities for expansion). The good – the interest, the value[viii] – would be either found in Natural Law, in Human Rights, or in the Constitution (among other suggestions), and a positive value would be granted to it by the legislator, who, judging it relevant, would grant it the protection of the law (legal good)[ix]. The scope of criminal law protection would include those legal goods that are so “fundamental to community life and the free development of the person”[x] to the extent that the state (alongside each of the subjects entitled to that interest) assumes a specific interest in ensuring the “undisturbed enjoyment [of the good] by each person who possesses it”[xi].

The harm principle

The Anglo-American approach to defining the extent of the state’s legitimate criminal intervention emerged in the form of the harm principle. Historically, it first appeared in J. S. Mill’s work On Liberty, but was refined by Joel Feinberg in a four-volume analysis of the limits of criminal law.

Mill states, in the first version of this principle, that the only purpose for which the state could employ criminal law, thereby restricting the freedom of individuals, would be to prevent harm to others[xii]. Feinberg later introduces what he calls “mediating principles”: considerations, beyond harm, that must also be taken into account when deciding whether to adopt criminal law. According to this author, criminalisation would only be legitimate if, in addition to preventing harm, the use of criminal law were the most effective way to prevent it, and if this intervention were carried out at minimal cost to other values[xiii] (considerations inherent to the principles of proportionality and necessity). Therefore, his stance is less extreme than that of his predecessor, as he considers that preventing harm to others is a good reason for adopting criminal law, but not the sole reason.

Unlike his predecessor, Feinberg attempted to define ‘harm’, identifying two possible meanings: one non-normative and the other normative. The first would correspond to a mere setback to a given interest, while the second would correspond to the unlawful (improper, incorrect) nature of the action. According to this principle, criminalisation would thus only be legitimate when the conduct to be criminalised caused harm to another; this ‘other’ being only the individual, not the community.

Although this principle has some concrete advantages, such as allowing for objective consideration of the damage caused by the crime and the damage resulting from criminalisation, it has also been criticised. Between the vagueness of its content, which can legitimise criminal expansionism, the difficulty of determining the individual or social sphere of the damage, and the use of the normative meaning of damage – criticisms that can be overcome in some way[xiv] – lies the most significant weakness of this principle: the fact that it focuses the legitimacy of criminalisation exclusively on the damage caused to the individual. This precludes the protection of collective interests (e.g. the environment) by criminal law, meaning they are referred to the sphere of civil or administrative law only. However, given that some of these interests are almost exclusively harmed by legal persons, protection through administrative law would be manifestly insufficient, as it could lead to the advantages of adopting the prohibited conduct being weighed against the (financial) disadvantages of paying the respective fine.

Thus, although some criticism is also directed at the legal good, the truth is that it proves to be a more complete criterion – and also more reviewable, given its external reference – than the harm principle, even though the latter focuses on actual damage to the individual (thus avoiding, at least in part, crimes of danger), and therefore has greater limiting potential.

Legal moralism

Another proposal, emerging also in the Anglo-American sphere, was to use legal moralism as a criterion for determining which types of conduct should be subject to the repressive power of criminal law. This philosophical trend in criminal law sparked a fierce debate with liberalism, which sits at the opposite end of the spectrum to moralism.

The crux of the problem with moralism lies not in determining that criminally relevant behaviours are (also) immoral – few would dispute this – but in asserting that the mere fact that a behaviour is considered immoral is sufficient to subject it to the coercive sanctions of criminal law. If it is sufficient that the conduct is contrary to a given social morality, then it is unnecessary for it to cause harm or offend any interest; it is sufficient for it to be contra bonos mores et decorum.

Lord Devlin is perhaps the best-known proponent of this theory (at least in its original version), having vehemently opposed the liberal function attributed to criminal law by Hart, one of the authors of the Wolfenden Report[xv]. The purpose of this report was to suggest relevant changes to the existing legislation on homosexuality and prostitution offences. In his lecture, Lord Devlin concludes that ‘the suppression of vice is as much the business of the law as the suppression of subversive activities’[xvi].

He argues, in support of his position, that a community is also defined by the shared ideas of its members about how they should behave and conduct their lives. These aspects would constitute the morality of that society. Therefore, integrating into society would imply that its individual members accept the shared morality and conform to behaviours considered correct or incorrect/moral or immoral by the ‘reasonable’ or ‘sensible man’ (not the majority of citizens), thus making it legitimate to resort to criminal law to prevent the disintegration of that society (in essence, to defend its existence)[xvii].

Following the academic debate with liberals and the reforms in the field of sexual criminal law to which they gave rise (effectively decriminalising conduct that should not be subject to state interference, such as homosexuality, prostitution or pornography), it would appear that moralism was definitively outdated as a proposal for defining the legitimate scope of criminal law. This is not only because the arguments in favour of moralism contradict the most fundamental concept of freedom recognised for human beings today, but also because, in truth, moralism effectively limit the punitive power of the state; rather, it justifies any criminalisation, provided that it is based on a supposed social morality.

However, it should be noted that new proposals in defence of moralism are emerging[xviii].  Nevertheless, the function of criminal law cannot be to defend any particular morality, even if the judgements of the legal and moral orders may coincide in certain circumstances. Moralism becomes definitively indefensible as a justification for criminal intervention when we move beyond the purely national sphere of action (where the existence of a homogeneous community is already questionable, but still easier to identify) and focus on the pluralistic international or European society. In this case, where differences are accentuated, how can we defend a (certain) morality that is considered sufficient to trigger punishment by the entity holding ius puniendi? For these reasons, moralism must (continue to) be dismissed outright as a justification for the exercise of punitive power.

Other intervening principles

In order for a criminal law to be legitimately adopted, it must not only aim to protect a fundamental interest (thereby having legitimate material content), but also respect other principles that are particularly relevant in the context of criminalising conduct. These include the principles of subsidiarity (which includes the ultima ratio principle) and proportionality, among others.

While these principles take on a slightly different meaning depending on the context in which they are applied (i.e. whether they are applied exclusively in a national or transnational context), they essentially determine that recourse to criminal law should be the state’s last resort – when no other branch of law is capable of adequately protecting the interest in question – and that such criminal intervention should not exceed what is strictly necessary for that protection. This is because excessive intervention in the fundamental rights of individuals (especially their freedom) is not legitimate when a lesser intervention would suffice to safeguard the interest in question[xix].

The option for the legal good and its consequences

In several legal systems, the option has been made to adopt the concept of legal good as the critical standard of criminal law. In these systems, a clear focus on a trans-systemic concept of the legal good can be seen[xx]. This means that the legal good worthy of protection is sought outside the criminal system (transcendent reference), yet still within the legal system as a whole (immanent reference). These external limits to the legal good can be found precisely in the Constitution, the fundamental law that regulates the most important aspects of society.

The selection of this reference point, as well as the choice of interests deserving of criminal protection, is based, first and foremost, on the circumstances required for restrictions on fundamental rights to be considered acceptable. Returning to the Portuguese example, Article 18(2) of the Constitution of the Portuguese Republic (CPR) states that “The law may only restrict rights, freedoms and guarantees in cases expressly provided for in the Constitution, and restrictions must be limited to what is necessary to safeguard other constitutionally protected rights or interests” (emphasis added; own translation). It follows, precisely, that if criminal sanctions seriously restrict constitutionally recognised fundamental rights (such as the right to liberty and the right to private property, which are respectively restricted by imprisonment and fines), then they can only legitimately do so to protect other interests recognised in the Constitution[xxi]. Recourse to the Constitution is also justified when identifying criminally relevant legal interests since it represents the “basic content of the social agreement”[xxii] governing the community in question. Thus, a law of lower hierarchy should not override the assessment inherent in the Constitution. Criminal law, thus defined in its content, demonstrates an ability to adapt to new social demands (provided these are reflected in the fundamental law), but it is also inherently restricted in terms of what can legitimately be criminalised, since a fleeting interest, that is not yet enshrined in the constitution, cannot be subject to the protection of criminal law.

The function attributed to criminal law, its raison d’être, is thus the “subsidiary (or ultima ratio) protection of legal goods with criminal dignity”, whereby “legal good may be defined as the expression of an interest, of the person or the community, in the maintenance or integrity of a certain state, object or asset that is socially relevant in itself and therefore legally recognised as valuable”[xxiii]. This enables a rational and reviewable limitation of state legislative action in the field of criminal law, legitimising the restriction of fundamental rights resulting from the sanction, since the unlawful act committed also jeopardised something valuable to the community (without which the community cannot survive in the same way). This is, in essence, the basis of the power to punish.

Some consequences arise from this notion of criminal legal goods, including some aspects (which sometimes tempt criminal legislators) that are ruled out by such a delimited function. Firstly, unlike the harm principle, for example, it should be mentioned that the concept of legal goods allows for the defence of collective goods. This principle is not only personal anchored in personal interests, since human beings are not only individuals but also social beings; in both dimensions, they need basic conditions for the development and enjoyment of each of their spheres of action. In this sense, the principle of the exclusive protection of legal goods is more comprehensive.

Conversely, it plays a significant negative role[xxiv] in determining what cannot be legitimately included in criminal law. Pure moral violations, ideological impositions or values of a purely administrative nature (resulting from mere state policy) will not be included in the concept of legal goods, due to the absence of a genuine fundamental interest of the individual or society (and consequently, the absence of a legitimate legal good). This function culminates in what is perhaps its most relevant consequence: the removal of criminal law that does not protect a legal good.

If submitted for review (namely, by the Constitutional Court), a criminal law that has not yet been adopted and is not based on a legal good should not be adopted; if already adopted, it should be removed from the legal system, as it is materially unconstitutional. This will be one of the most important functions entrusted to the Constitutional Court, even though they do not always apply the concept of legal goods as an insurmountable limit to the state’s ius puniendi consistently.

Whether the same function can be attributed to criminal law stemming from transnational spheres is another question, as in these settings there are other concerns that may lead to a somewhat differentiated solution[xxv]. These include, for example, the need to account for different national identities, constitutional standards and rules, which demand a more flexible approach. However, a substantive limit on the legislative power to criminalise conduct should always be employed, to ensure that restrictions on individuals’ fundamental rights legitimate – to protect other relevant interests, thus limiting the illegitimate overexpansion of criminal law or its use as a mere threat to guarantee individuals comply with public policies of various kinds.

This text aims to consider the legitimacy of criminal law in general. For a more detailed analysis, please refer to the list of sources. The text retains an informal tone, as it is based on a presentation delivered at a multidisciplinary international congress in November 2025.

Endnotes

[i] With Klabbers, Jan, “Setting the Scene”, in: Jan Klabbers; Anne Peters e Geir Ulfstein (Eds.), The Constitutionalization of International Law, Oxford, Oxford University Press, 2011, p. 39 f.

[ii] Legitimacy as a pre-condition of this conviction can be explored in more detail in Hurd, Ian, After Anarchy: Legitimacy and Power in the United Nations Security Council, Princeton, Princeton University Press, 2007, p. 7 f.

[iii] Birnbaum, Johann, “Über das Erforderniß einer Rechtsgutverletzung zum Begriff des Verbrechens, mit besonderer Rücksicht auf den Begriff der Ehrenkränkung”, in: Archiv des Criminalrechts, Neue Folge, 1834.

[iv] Feuerbach, P. J. Anselm, Lehrbuch des gemeinen in Deutschland gültigen peinlichen Rechts, 14. Auflage, Giessen, Georg Friedrich Heyer’s Verlag, 1847.

[v] Binding, Karl, Die Normen und ihre Übertretung, Erster Band, Leipzig, Verlag von Wilhelm Engelmann, 1872.

[vi] Liszt, Franz von, “Der Begriff des Rechtsgutes im Strafrecht und in der Encyklopädie der Rechtswissenchaft“, in Zeitschrift für die gesamte Strafrechtswissenschaft, Vol. 8, 1888.

[vii] Even though there are other jurisdictions where the same limit is applied, such as Germany or Italy.

[viii] In detail, Cardoso, Raquel, As Funções do Direito Penal Europeu e a Legitimidade da Criminalização. Entre o harm principle e a protecção de bens jurídicos, Coimbra, Almedina, p. 38 f.

[ix] Fiolka, Gerhard, Das Rechtsgut. Strafgesetz versus Kriminalpolitik, dargestellt am Beispiel des allgemeinen Teil des schweizerischen Strafgesetzbuches, des Strassenverkehrsgesetzes (SVG) und des Betäubungsmittelgesetzes (BetmG), Basel, Helbing & Lichtenhahn, 2006, p. 155.

[x] Dias, Jorge de Figueiredo, Temas Básicos da Doutrina Penal, Coimbra, Coimbra Editora, 2001, p. 157 (free translation).

[xi] Moro, Aldo, Lezioni di Istituzioni di diritto e procedura penale, Bari, Cacucci Editore, 2005, p. 203 (free translation).

[xii] Mill, John Stuart, On Liberty (1859), Kitchener, Batoche Books, 2001.

[xiii] Feinberg, Joel, Harm to Others. The Moral Limits of the Criminal Law, Vol. 1, New York, Oxford University Press, 1984.

[xiv] Cardoso, Raquel, As Funções do Direito Penal Europeu e a Legitimidade da Criminalização. Entre o harm principle e a protecção de bens jurídicos, Coimbra, Almedina, p. 108 f.

[xv] Report of the Committee on Homosexual Offences and Prostitution, Her Majesty’s Stationary Office, 1957 (Wolfenden Report).

[xvi] Devlin, Patrick, “The Enforcement of Morals”, Maccabaean Lecture in Jurisprudence, Oxford University Press, 1959, p. 141.

[xvii] Ibid., p. 136 ss.

[xviii] In analysis, Cardoso, Raquel, “Moralismo legal: uma “nova” opção para o Direito Penal?” (forthcoming).

[xix] Cardoso, Raquel, “Legitimidade e necessidade no Direito Penal (Europeu). Um diálogo, dois discursos”, in Revista Portuguesa de Ciência Criminal, Ano 33, nº 2, 2023, p. 235 ss.

[xx] For a definition, Dias, Jorge de Figueiredo, Direito Penal. Parte Geral. Tomo I. Questões Fundamentais. A Doutrina Geral do Crime, 3ª Edição, Coimbra, Gestlegal, 2019, p. 132 ss.

[xxi] Durán Migliardi, Mario, “Constitución y legitimación de la pena. Apuntes teleológicos sobre el rol de la Constitución en el sistema penal”, in Politica Criminal, Vol. 6, Nº 11, Junio 2011, p. 146.

[xxii] Díez Ripollés, José, “La contextualización del bien jurídico protegido en un Derecho Penal garantista”, in Ouviña, Guillermo (etc.), Teorías actuales en el Derecho Penal, Buenos Aires, Ad Hoc, 1998, p. 432.

[xxiii] Both quotes from Dias, Jorge de Figueiredo, Direito Penal. Parte Geral. Tomo I. Questões Fundamentais. A Doutrina Geral do Crime, 3ª Edição, Coimbra, Gestlegal, 2019, p. 129-130 (own translation).

[xxiv] Ibid., p. 141 f.

[xxv] Cardoso, Raquel, As Funções do Direito Penal Europeu e a Legitimidade da Criminalização. Entre o harm principle e a protecção de bens jurídicos, Coimbra, Almedina, p. 395 f.