International Journal for the Semiotics of Law - Revue internationale de Sémiotique juridique

, Volume 22, Issue 2, pp 223–243

Individual Conceptual Structure and Legal Experts’ Efficient Communication


    • Aarhus School of BusinessUniversity of Aarhus

DOI: 10.1007/s11196-009-9104-x

Cite this article as:
Engberg, J. Int J Semiot Law (2009) 22: 223. doi:10.1007/s11196-009-9104-x


The article investigates characteristics of legal concepts as found in academic articles, focusing upon the knowledge base of legal experts. It is a cognitively oriented study of one of the semiotic basics of communication for academic legal purposes. The purpose is to study the structure of knowledge elements connected to the concept of “Criminal liability of corporations” from US law in and across individual experts in order to look for individual differences and similarities. The central concern is to investigate the conditions for the observable efficiency of semiosis in academic discourse. In a first basic section I discuss aspects relevant for a cognitively oriented study of academic discourse. The empirical part of the article consists of an analysis of text passages from two articles in American law journals. The results of the study support the assumption that high efficiency and precision of semiosis is due rather to the use of specific cognitive processing skills than to total identity of cognitive structures across individual experts.


In this article, I want to look at characteristics of legal concepts with a focus upon the structure and content of the knowledge base that individual legal experts use when they communicate about their field on the basis of a study of written text material. The underlying research interest is to investigate the conditional relation between the human cognitive system and the observable high efficiency and precision of expert communication, on the one hand, and between aspects of this relation and argumentative writing in the legal field, on the other. In other words: How is high communicative precision possible under the conditions of human cognition and what are the relations between these conditions and the way legal experts write their articles? The article is consequently oriented towards cognitive and knowledge-oriented aspects of legal communication and semiosis.

A major source of inspiration for choosing this approach to the study of legal communication has been work by Roelcke [35]. He claims that a new type of studies of LSP has emerged since the beginning of the 1990s, the members of which are subscribed to what he calls the cognitive-linguistic function model (Kognitionslinguistisches Funktionsmodell). In such studies, focus is on the communicating parties and their specialised mental capacities and contexts. Studies following this approach may be seen as contributions to set up the “missing link” between the linguistic system and the specialised context. They intend to be explanatory in the way that they explain not only why communication is specialised, but also how the specific and observable characteristics of the linguistic specialisation emerge. They are so to speak concerned not only with the influential factors, but also with the mechanics lying behind the specialised communication [36].1 The present study intends to be an example of this type of study. It has its focus on assessing characteristics of the legal (=specialised) knowledge on which individual experts draw when participating in specialised academic communication in their field. Studying this object is also a study of the stored specialised meanings of the terms used in the legal field. It is thus intended as a study of cognitive factors in semiosis, but not a traditional study of semiotic elements. I want to study specialised meaning as seen not from a text perspective (although empirically dependent on texts), but from the perspective of knowledge bases: Looking at meaning from a text perspective focus would be on the meaning of a word in its (actual) textual and wider context; looking at meaning from a knowledge perspective, instead, the focus is on the cognitive base of the experts involved in the communication. To underline the difference in perspective, in this article I will be talking about meaning and meaning making, when I consider the linguistic parts of the process of establishing the meaning of communicative contributions; and when the stored knowledge base is meant, about knowledge, concepts and knowledge construction. I will start out by discussing some of the consequences of focusing on the human cognitive system for the study of legal communication as an instance of specialised communication in Sects. 24. In these sections, I will concentrate upon three questions connected to the so-called cognitive-linguistic function model: What is the central object of study (2)? What semantic assumptions and approaches are relevant (3)? How can precision and efficiency be achieved given the affordances of the cognitive system (4)? In Sect. 5 I will present a methodology that meets the requirements laid out in the previous section. It is applied in the study of the concept of “Criminal Liability of Corporations” in US law as represented in articles in law journals reported in Sect. 67. Finally, Sect. 8 concludes the article and briefly discusses consequences for our conceptualisation of specialised knowledge in connection with meaning in specialised communication.

What is the Central Object of Study?

Following a cognitive approach to the study of knowledge and meaning in legal communication central objects are the complex concepts of experts as they are represented, among other places, in written texts and consequently made mutually manifest to the communicative parties. Investigating the complex concepts of legal experts, focus may be either on the concepts of the individuals participating in the communicative event, on the one hand, or on the totality of knowledge of individual experts in the field, on the other.

The latter has been the typical object of study for a traditional knowledge-oriented LSP discipline like terminology. Interest has been on concepts and the structural relations between them as they may be assessed when looking across a number of individual instances. This set-up has been spurred on by the traditional assumption that a disciplinary field consists of an amount of knowledge shared by the individual experts constituting the disciplinary field. In the centre of interest under such a traditional perspective we find the elements of knowledge that the experts agree upon in communication, toning down the differences. This is sensible and relevant, if we want to describe the knowledge of a discipline, i.e. of a supra-individual entity, like when we for example compile dictionaries or set up text books with the central knowledge of a discipline.

However, such a more structuralistic approach has the disadvantage that it will tend to ignore the mechanisms of development in disciplinary knowledge. It is thus not optimal from the point of view of studying actual semiosis. For developments actually take place at the margins of the agreed knowledge, in discussions about (parts of) the knowledge and as a result of competing conceptualisations of a specific concept. In order to better grasp such developments also in the study of specialised communication, the concept from Michael Halliday’s work on Social Semiotics [23, 47] has been applied (e.g. [21]). In this approach, the knowledge underlying meaning-making processes (i.e. the semiotic potential; [47, p. 4]) is seen as what I would term a socially inspired entity. The idea is that the knowledge base of an individual is influenced and shaped by the interaction with others and by the social relations and conditions under which these interactions are performed [23, p. 6, 47, p. 26]. Knowledge and meaning are thus not stable entities, but dependent on influences from social factors that change with time.

This type of approach solves the problem of accounting for the constant development in knowledge and meaning observable in actual specialised communication. However, studies in Social Semiotics are normally looking at the shared knowledge of the members of the investigated social group. They also tend to background the individual processes and the consequent differences, as their interest is in the socially based meaning potential of the investigated semiotic resources rather than in the actual individual knowledge bases [47, p. 4]. So the approach does not traditionally focus upon the individual experts, and consequently they do not fulfil the above-mentioned need for such an alternative focus.

I take this to be an important desideratum in the field of studies of specialised knowledge and meaning also in law. I follow here the argumentation by Harder [22], who with reference to Searle [38] states about language and meaning that they are inherently two-sided: they are social facts, but at the same time grounded in actual mental representations of individuals: “To begin with, social facts are only possible by virtue of the existence of individual minds capable of representing such social facts to themselves” [22, p. 206]. Concentrating on the social aspect of such facts, emphasising shared aspects of in our case the knowledge like it is seen in the above-mentioned approaches, moves interesting aspects of the cognitive mechanism and the knowledge basis underlying the process of making specialised meaning to the background. The alternative approach advocated here intends to foreground some of these aspects, i.e. the influence of the individual processing on the socially induced generation of knowledge. Therefore, in order to have a more complete picture of the semiotic process in specialised legal settings it is necessary that the traditional approaches are supplemented by an approach focusing on the individual expert, his or her stock of concepts, and the interaction of the concepts of individual experts in concrete situations. In the remainder of this article, I will discuss some of the basic elements of such an approach and suggest a relevant methodology. And in my conclusion I will briefly discuss some of the implications of such an approach outside the realms of academic description of linguistic meaning.

I am of course not the first person in the field to try to fill the gap. Relevant work, which has functioned as inspiration for me, is, e.g. Ahmad and Musacchio [1], a corpus study of how the Italian physicist Fermi together with his colleagues constructed an Italian vocabulary for the field of nuclear physics in order to express their newly developed ideas, Bazerman [3] on the rhetorical presentation of light by the inventor and producer of the light bulb, Thomas Edison, and Laurén [26] and Nordman [31], both on the idiolects of individual Swedish sociologists.2

Focusing upon the knowledge of the individual expert has an important consequence in the way that it brings up the question, what this knowledge actually looks like in its individuality: Is the knowledge of the individual expert an exact mirror of the knowledge of the field, in which case experts share concepts; or is the disciplinary knowledge of individuals exactly individualised, i.e. does it have an overlap with the concepts of others, but without being an exact mirror? Section 4 below reports an empirical study with this focus. However, before we get to that, I would like to discuss the usefulness of different approaches from semantics to the suggested alternative approach.

What Semantic Assumptions and Approaches are Relevant?

Traditionally in the field of LSP studies focusing upon specialised meaning and especially in the field of terminology, the underlying assumption about efficient and precise specialised communication has been that it functions optimally when all participants in the communication have identical concepts to draw upon when communicating (see e.g. [2, p. 24]). Consequently, the interest of terminology studies has traditionally been in standardisation, i.e. in setting up systems of concepts that experts could take over in order for them to have an identical basis for their communication.3 We find the same type of basic assumptions underlying cognitively oriented approaches to general semantics that suggest systems of semantic universals (e.g. [48]) or postulate mutual mental resources like a language-independent “language of thought” (e.g. [13]). The traditional terminology approach and the mentioned approaches to general semantics, for which I will in the following use the cover term “code-oriented approaches”, share at least two interconnected assumptions about the functioning of the cognitive system, which may be deduced from their respective argumentation:
  • The cognitive system infallibly yields correct interpretations when understanding words (or should optimally do so);

  • The cognitive system does so in isolation (or should optimally do so).

The most efficient way to meet the requirements of such assumptions is to stretch the code aspect of language as far as possible and presuppose an identical conceptual basis in all individuals participating in communication.

However, the two assumptions are challenged by empirical results from work within different approaches to semantics in a wider sense as well as within psychology. A few examples should suffice here. From the field of semantics, Schroeder [37] challenges especially the infallibility assumption in an article critical of Fodor’s above-mentioned approach. He states that the degree of actual misunderstanding found in communication is at odds with the idea that understanding in communication should be dependent on identity of concepts in the people participating in communication. As an alternative, he sets up a methodology to be used in order to measure whether the concepts of two persons are sufficiently similar for them to engage in a process of agreeing or disagreeing about the meaning of words used in communication [37, pp. 75–80]. Also from the field of semantics, the idea of the identical conceptual basis among participants in the communicative event is challenged by Sperber and Wilson when developing Relevance Theory:

Human beings somehow manage to communicate in situations where a great deal can be assumed about what is manifest to others, a lot can be assumed about what is mutually manifest to themselves and others, but nothing can be assumed to be truly mutually known or assumed [41, p. 45].

What both works hint at is the fact that the process of making meaning in communicative situations is in fact fairly risky business, and that it does often go at least mildly wrong, especially in the first stages of interaction. This challenge of the infallibility assumption then leads to a challenge of the isolation assumption: Code-oriented approaches give the individual linguistic system the primary responsibility for securing correct understanding, whereas approaches built on similarity of concepts lay more responsibility on the communicative interaction of the participants—and therefore leave more room for the actually observable misunderstandings.

From the point of view of psychology and the study of concepts in cognitive systems, Millikan [28] also rejects the necessity of the infallible system in order to account for actual communication. In her view, the primary function of concepts (units of stored knowledge) is to “facilitate information gathering and use for an organism navigating in a changing and cluttered environment” [28, p. 59].4 Concepts like the concept cat thus serve primarily as instruments for recognising encountered animals as instances of a specific kind, in order for the human cognitive system to be able to learn from individual encounters with such animals and to carry over knowledge from one encounter with instances of the kind cat to another. In order to serve this purpose, she argues, it is more efficient to have a concept structured as a collection of indicators used for recognising instances of the kind than to have concepts in the form of propositionally structured definitions. The latter type of concepts would enable the cognitive system to infallibly deduce whether an encountered object belongs to a specific kind. But it would in many cases filter out elements that actually belong to a kind (e.g. if the basic definition to be used for the deduction is not yet fully developed). Therefore, she states that concepts structured around appearances of instances of a kind, despite being fallible, are far more efficient in practice. From a psychological point of view such structures are consequently more likely in the light of the purposes to be fulfilled and the observable behaviour [28, p. 59].

And finally, studies of human memory also challenge the assumption of infallibility. This is due to the fact that the assumption presupposes stable memory structures in order to keep concepts identical and thus secure infallibility. However, such memory structures are difficult to find. Empirical studies indicate that memory is more of a function than a stable structure. Distinguishing between episodic and semantic memory is a case in point that may function as an example here. Originally, these two types of memory were seen as structurally different [45, 46]. Empirical work indicates, however, that it is more adequate to see these two types of memory as merely different constructive functions [11, 12]. As an example, McCabe and Balota [27] show that memory results, in the concrete case: which recognised elements are seen as being remembered (=episodic memory) as opposed to just known (=semantic memory), are dependent on the context in which the elements are learnt or presented for remembering as well as on the context of the task in which the process of remembering takes place. The results show that the function of memory can hardly consist in a mere copying of structures stored in the brain. Instead, as a function memory is rather a reconstructive process on the basis of a stored substrate.5 And furthermore this function is dependent on the type of information needed in communication (information about specific encounters with an item (episodic memory) versus information about the item in general (semantic memory)) and on the type of task the function is involved in.

This state of affairs is relevant for the discussion of the assumption of infallibility of the cognitive system concerning semantics. For if memory is more of a function and if this function is recreative and context dependent, then these system conditions render it fairly unlikely that the cognitive system should be able to infallibly find the correct interpretation on the basis of a mutual set of universals or agreed concepts.

In conclusion, the infallibility assumption is challenged by what we observe in actual communication as well as by results from psychology about the characteristics of the cognitive system. As a consequence, the assumption that the cognitive system works or must work in isolation is also challenged, as one way of actually mitigating the consequences of fallibility is to interact with other cognitive systems. At the same time, Fodor and Lepore argue for the necessity of accepting the original assumptions and the consequent presumption of inter-individual conceptual identity in communication in order for semantics to be able to do the things a semantic theory should do [15, p. 383]. This is a dilemma. How do we as interested in meaning in academic legal texts navigate in this field? By finding out at what level our interests lie. Interestingly enough, Fodor [14, pp. 38–39, 78] also states that his approach is not feasible at more complex cognitive levels, although he thinks it is the optimal approach to model what goes on in linguistic cognition in a narrow sense. I will not enter into a discussion of where the exact borderline is to be drawn for the efficiency of code-oriented approaches, or whether the above-mentioned assumptions are actually necessary for semantics to fulfil its functions, as it would go beyond the scope of this article. However, it is obvious that at the higher order of cognitive processes relevant when looking at the knowledge of legal experts the actual cognitive systems of the experts do not function in isolation, but in interaction with the context and in interaction with other cognitive systems (discoursive grounding, [39, pp. 244–248]; and the experts demonstrably do not fully agree upon the content of the concepts that are subject to the communicative event.6 So they do not infallibly reach the same interpretations in communication. Instead, in the academic legal communication process there is a constant interaction going on with other legal experts, explicitly and implicitly: Sender and receiver draw upon previous arguments and positions held by the participants or by others, upon previous experiences in similar situations, and upon relevant knowledge from other situations that is associated with the situation at hand and is thus deemed relevant by the individuals. And these are the assumptions a relevant approach has to be based upon.

So, a cognitively oriented approach to the study of LSP texts should look at meaning as a procedural entity and at the knowledge basis for this process. But it should at the same time focus on the potentially messy reality of instances of specialised communication in academic settings that emerges because participants have somewhat different and maybe not well-structured knowledge bases. And this reality is situated at a level of cognitive complexity that makes code-oriented approaches like the presented ones from general semantics, but also traditional terminology, irrelevant, even though the general semantics approaches may be good approximations of what goes on at more basic cognitive levels.7

I hope to have argued convincingly for choosing an approach based on fallibility, cooperation, and concept similarity for investigating semiosis on the basis of specialised knowledge in academic legal communication. Choosing such an approach has especially one consequence that I would like to dwell upon, before presenting in Sect. 5 the methodological considerations, on which the empirical study in Sect. 6 is based, viz. the necessity of monitoring the Other in a communicative event, when intending to make meaning on the basis of the Other’s communicative contribution.

How May Precision and Efficiency be Achieved?

The question in the section heading becomes relevant because in the previous section we argued that we cannot (as this has traditionally been done in the field of LSP studies) presuppose identical, but only at best similar conceptual structures across individuals in LSP communication. So the observable precision and efficiency in LSP communication must be due to a different cause. In this connection, I suggest that the concept of Theory of Mind as introduced by Premack and Woodruff [33] may be relevant as an element of explanation. In their use of the phrase, for an individual to have a “Theory of Mind” means that “the individual imputes mental states to himself and to others (either to conspecifics or to other species as well)” [33, p. 515]. In later work, Theory of Mind has been shown to be a natural part of normally functioning human cognition.

The Theory-of-Mind ability of humans allows us to form and hold representations of own and foreign concepts at the same time. A person with this ability knows, e.g. that he or she has one concept connected to a word, whereas others may have a completely or a somewhat different concept (different or similar concept, respectively) [32, pp. 59–61, 42, pp. 10–13]. Introducing this capacity in the explanation solves one of the potential problems concerning the efficiency of expert communication. For it makes lack of identity of concepts unproblematic, as long as the participants are able to “tune in” on each other on the basis of a similarity of their respective concepts, their experience from previous communicative interactions, and the actual interaction with the other participants.

The aspect of Theory of Mind that I will concentrate upon here is the knowledge one participant in communication ascribes to other participants when understanding them (semantic aspect) rather than the attribution of communicative intentions, ideological standpoint, etc. (pragmatic aspect).8 On the basis of what we have seen so far, I assume that efficient and precise understanding in a specific communicative interaction is generally possible, if at least one of two requirements are met: (1) The concepts of the parties are sufficiently similar for the parties to enter into a process of negotiating (=making) actual meaning (in which case they may, but must not end up agreeing about the meaning of a word); or (2) the concepts are not similar, but the parties’ mind-reading capacities and stored knowledge about what concepts would regularly be evoked by other experts are sufficiently developed for them to be able to infer what the other person is thinking in the specific situation (in which case they do not agree with but can at least understand what the Other says). In the case of experts communicating efficiently, this means that the experts must be experts in the same field and that, if their concepts are not identical, they at least know that someone in the field may hold the concept that the other person holds.

From these requirements it is obvious that agreement on the structure or profile of a specific concept and its constituents is not necessary in order to get the communicative interaction and the subsequent process of understanding running in an efficient way. What it takes is that the participants know what alternatives are normally held within the field, and that the participants stay inside the range of normally held conceptualisations. This makes it possible to have an efficient Theory of Mind about the concept of the other participant and to be aware that identity of knowledge may not exist. Full identity of concepts is thus not necessary in order to achieve the kind of efficient understanding that may be observed in actual expert communication.

To sum up the remarks in this Sect. 24: I take it to be relevant to look at the semiotic process in legal communication from the point of view of the individual knowledge bases of the participants, in order to have a fuller picture of meanings in legal communication. However, this means that we must leave the idea of identical knowledge bases as the reason for efficient communication. Instead, the concept of Theory of Mind set up on the basis of potentially different knowledge bases offers itself as a relevant alternative. In the empirical part of this article (Sect. 6) I will investigate especially to what extent the results of the study may be said to back up the assumption of an important role of Theory of Mind in expert communication. To this end I will apply the method of semantic networks, which I will present in the following section.

Semantic Networks as an Analytical Tool

In order to fulfil the requirements set up in the previous sections, a relevant approach should ideally intend to document the messy and individualised picture that we find when studying actual academic communication. Corpus studies in the traditional sense with focus on the general trends are thus not the ideal solution for the reasons stated above. Instead, we need to focus upon the individual expert’s individualised communicative efforts, i.e. the knowledge reflected in the individual text.

In order to meet these requirements, I have chosen to work with the methodology of semantic networks suggested by Gerzymisch-Arbogast and Mudersbach as presented first in [29] (cf. also [1619]. The methodology is based on the assumptions of dependency grammar and thus takes the sentence and its internal syntactical relations as input for constructing models of sentence and text meaning. The focus is upon the graphical representation of arguments and relators and their relations.9 According to the needs of the investigation a relevant level of detail may be chosen. In this article, I work with a restricted version of the methodology as developed in previous work [8, 9]. In this version I restrict myself to representing associative relations between arguments (prototype:noun) and relators (prototype:verb) in the text without indicating the specific kind of relation between them. As an example, Fig. 1 shows the graphical representation of the sentence “A concept of law is a knowledge structure that is placed in the mental system”.
Fig. 1

Semantic network of example sentence

As is visible, relations are represented between arguments and relators (e.g. between concept and be), between arguments (between concept and law), and between relators (in the case of subordinate clauses, here between be and place). Elements that reoccur in the linguistic structure (like the element knowledge structure that occurs in the superordinate as well as the subordinate clause) are not repeated, but are connected to the substructures they are part of.10

In the following empirical study I have applied the methodology in order to show the knowledge elements represented in the investigated text excerpts and the relations between them.

Empirical study—Criminal Liability of Corporations

The concept to be studied in this empirical investigation is “Criminal Liability of Corporations”. The following definitions should give an idea of the profile of the concept:

liability, n. 1. The quality or state of being legally obligated or accountable; legal responsibility to another or to society, enforceable by civil remedy or criminal punishment

criminal, adj. 1. Having the character of a crime; in the nature of a crime

corporation, n. An entity (usu. a business) having authority under law to act as a single person distinct from the shareholders who own it and having rights to issue stock and exist indefinitely; a group or succession of persons established in accordance with legal rules into a legal or juristic person that has legal personality distinct from the natural persons who make it up, exists indefinitely apart from them, and has the legal powers that its constitution gives it. (Black’s Law Dictionary, 8th ed. 2004)

The concept “Criminal Liability of Corporations” refers to legal obligations of a legal person (constituted by a number of natural persons) with a personality distinct from these natural persons. And the obligations are rooted in criminal behaviour. The interesting thing about this concept is that its content, i.e. the knowledge connected to it, has changed fundamentally over time. Traditionally (before, say, 1850) corporations were generally seen as precluded from being criminally liable as such, as they had no “soul” and therefore could not have a so-called mens rea (sc. a guilty mind or criminal intent, Black’s Law Dictionary, 8th edition, 2004). Instead, the persons in the corporation actually acting criminally (and not the corporation employing them) would be criminally liable for their own actions. However, this state of affairs was challenged by the emerging view that companies should be prevented from circumventing their responsibility by using employers as scapegoats. Instead they should be liable in their own right, just as the corporations were able to achieve rights through actions by its employees. This is generally seen as a kind of vicarious liability:

Vicarious liability Liability that a supervisory party (such as an employer) bears for the actionable conduct of a subordinate or associate (such as an employee) based on the relationship between the two parties. (Black’s Law Dictionary, 8th ed., 2004)

The concept of corporate criminal liability is interesting for us, because it is an example of the not uncommon constellation that a legal concept is caught in the crossfire between two opposing ideals: the protection of the corporation against being liable for actions performed by others (here: its employees) versus the protection of society against the inability to try the proper culpable for a crime. The clash between these two ideals has been constant over time. The important issue is that US criminal law regularly applies the mens-rea requirement mentioned above as obligatory for someone to be criminally liable. And the problem is how to prove this in the case of a legal person without a mind of its own. In US-Law, two landmark cases may be used to show the development of the concept.

In 1909, the US Supreme Court decided that a railroad company was criminally liable for the actions of one of its employees, in casu an assistant traffic manager, who had given rebates to a customer, although such rebates were forbidden by statute. Thus, at least from then on corporations were to be held criminally liable for actions performed by employees within the scope of their employment and for the benefit of the corporation, although the corporation as such did not have the criminal intent. And in 1987, the US Court of Appeal held a bank liable for the criminal actions of their employees, although the bank had informed the employees not to perform the action (=fail to report transactions in currency above $10.000), and although not one employee had the relevant knowledge about the size of the transactions, but rather the collective of employees. Thus, at least from then on, a corporation may be held criminally liable also for expressly prohibited actions and on the basis of merely collective knowledge, as long as it has not taken sufficient measures to prevent violations of the prohibition. This last position has been modified in later decisions, but is in principle still the law in the US system.

From the perspective chosen in this article, i.e. looking at the process of meaning making from the point of view of the individual’s knowledge base as source for Theories of Mind, the treatment of an inherently challenged concept is especially interesting. A legal expert engaging in argumentative communication knows and may represent the two opposing ideals in his communicative contributions. But he may have to make up his mind as to which ideal is the most relevant in his mind and on the basis of his knowledge, depending on the requirements from the communicative situation.

The investigation reported in this article concentrates upon instances of textually reflected knowledge of two individual legal experts around the last of the landmark cases mentioned. I have chosen to look at articles from around a landmark case (1987–1994) because these cases typically function as triggers of discussion of challenged concepts like the one treated here. And I have chosen to look at articles in law journals, because they are the most likely fora for discussions about relevant characteristics of the concept. In journal articles, the legal expert can state in detail what the concept looks like for him and also discuss other assumptions than his own, if these are relevant. Consequently, if the Theory-of-Mind ability is of relevance in this kind of communication, it will be visible here.

The empirical study was carried out on two texts from a corpus of relevant texts. The texts for the corpus were found by searching for law journal texts in the legal document service Westlaw. I used the following search terms: VICARIOUS LIABILITY, CRIMINAL, CORPORATION and CORPORATE and limited the search to articles between 1987 and 1994. This search produced a total of 31 relevant articles (relevant: criminal liability of corporations is a central topic of the article), of which two were chosen for the study due to the fact that they show a relevant amount of agreement and disagreement to investigate empirically the points treated in the discussions in the previous sections about the characteristics of experts’ conceptualisations. In these two law journal articles, I have looked for central parts of the argumentation concerning the role of the concept of criminal intent in stating that a corporation is criminally liable. Four of these sections I have subjected to a semantic network analysis.

Empirical Analysis

The two authors agree that the mens rea requirement is relevant in connection with criminal liability of corporations, but they disagree as to the feasibility of applying the requirement when judging this type of cases.

The first article [4] treats a proposal to amend the Texas Business Corporation statutes in order to prevent a successor corporation from being held criminally liable after a merger for actions performed by a predecessor corporation. The author thus talks about a specific case of criminal liability, but in his remarks also treats the general aspect of the social stigmatization by criminal punishment and its viability. His basic position is that it is conceptually impossible to apply the concept of mens rea to corporations.

The second article [5] is equally sceptical concerning the traditional application of the mens rea requirement in connection with the concept of criminal liability of corporations. However, instead of rejecting the viability of applying the concept to corporations altogether, she suggests evaluating cases according to a more justified scale and thus achieving a better operationalisation of the concept.

In the following, we will have a look at some examples of the argumentation by the two authors and thus try to get an insight into the individual knowledge structures of the authors. The first example is from Becker [4]:
  1. (1)

    Some current statutes state that a corporation can have a mens rea. This seems a far stretch of the imagination, but stating that a successor corporation has the mens rea to commit its predecessor’s crime is utterly preposterous. […] Even in the general corporate context, it has been argued that corporations cannot have a mens rea. As one academic put it, “The concept of wrongdoer is highly individualistic. It presupposes personal qualities: the capacity to have an intention and to choose”. This same author also stated that “the costs and burdens of reformation are placed on the organization, not because it has ‘done wrong,’ in either a literal or metaphorical sense, for it has neither an intention nor a will…”. These arguments have been rebutted with the argument that corporations are capable of intent in the form of corporate policy. [4, pp. 438–439]

On the basis of the example the semantic network in Fig. 2 may be set up on the basis of the principles outlined in Sect. 5.
Fig. 2

Semantic network of example 1 [4, pp. 438–439]

The semantic network shows what arguments and relators have been applied in the formulation of the argumentation. I see this type of semantic networks as a reflection of the elements of knowledge that are activated in the mind of the author in connection with his formulation of the argumentation and of the relations between them in his knowledge base. As this network has been activated in connection with discussing a specific aspect of the concept of criminal liability of corporation, viz. the mens-rea requirement, I take it that these elements and their relations may be said to be part of the knowledge connected to the concept in the mind of the author.

From our point of view, it is noticeable in this example that the activated knowledge is not just limited to some kind of monolithic definition or description of the concept. Instead, what we find are descriptions of mutually exclusive positions held by different experts explicitly ascribed to specific instances of experts. In four instances (the black relators) different positions are cited, and two different and mutually exclusive positions are held (“one academic argues…” versus “some current statutes state…” and “[someone] rebuts”). And in two instances the author explicitly has connected his personal opinion to the cited statements (greyish elements), in both cases evaluating the statements as being not credible or relevant. The interesting thing in our context is that this must mean that the individual expert’s knowledge connected to the concept of criminal liability of corporations contains not only elements that he believes to be true and correct. The knowledge base contains also elements held by others and refuted by the individual expert himself in the course of an evaluation process on the basis of his own cognitive processes. So what we see reflections of in the text is an individually profiled system of knowledge elements which is wider and less focused than the system of assumptions the expert himself holds to be true.

The second example is also from Becker [4]:
  1. (2)

    The following arguments have been offered in favour of corporate criminal liability. First, […].

    Conversely, the following arguments have been offered as arguments against holding corporations criminally liable. First, […]

    The problem with both sides of this analysis is that the commentators do not address the real issue: can a corporation have the guilty mind (mens rea) required for the committal of a crime or for the imposition of liability? No! It is that simple. [4, p. 469]

The semantic network of the example looks like the one in given in Fig. 3.
Fig. 3

Semantic network of example 2 [4, p. 469]

The network shows us a similar picture as before, with two opposing groups of arguments in the picture (black elements). However, this example is even more accentuated than example 1, as the author takes a view that actually refutes the relevance of the cited arguments (greyish elements). So the knowledge connected to the concept of criminal liability of corporations in the knowledge base of the legal expert writing this article must contain two groups of elements of which he is himself convinced that they are not relevant. Once again this is an example of an individually profiled set of knowledge elements which is wider and less focused than what the expert himself takes to be true.

Our third example comes from Bucy [5]:
  1. (3)

    With the rising prominence of corporate actors, our concept of intent must continue to evolve. Common to our past and present notions of intent is a focus on the individual actor. Our attempt to apply this notion of individual intent to corporate actors, however, has failed. To date, our approach has been to impute an individual actor’s criminal intent to the corporate actor. This Article suggests that this approach is not only inadequate, but also harmful because it erodes the power of the criminal law (p. 1107)

The semantic network of example 3 is given in Fig. 4:
Fig. 4

Semantic network of example 3 [5, p. 1107]

In this example, as opposed to what we saw above we do not have mutually exclusive or opposing knowledge elements present in the same section. But we do see a personal profiling consisting in a negative evaluation of the concept of intent that is postulated to be the widespread one. This profiling is visible in the greyish elements. First, the author emphasises that the existing concept must evolve, indicating that the existing concept is not adequate. And secondly, she connects the approach (the dotted line indicates that with approach she means the elements connected with the relator apply at the top of the network) with the negative elements harmful, inadequate and eroding and marks these connections as her own suggestions. She clearly has a personal (and negative) perspective on the subset of knowledge elements concerning the concept of criminal liability of corporations that she holds and attributes to others. So, again, if asked to come up with what is true and correct about the concept, she would hardly come up with all the elements shown in the network. They are part of her relevant system of knowledge elements, but not part of her own idea of what the concept (should) look(s) like.

The last example is also from Bucy [5]:
  1. (4)

    Using respondeat superior to prove and assess criminal liability of corporations essentially requires only proof that the corporation employed the miscreant actor. In this way, respondeat superior treats all corporations as if they were alike, and holds them all responsible for their agents, no matter what. Likewise, the MPC standard requires only proof that a high managerial agent committed the unlawful act, or supervised the miscreant agent. Although the MPC thus narrows the broad swath of the traditional respondeat superior standard, it still fails to sufficiently distinguish among corporations. Additionally, the MPC standard discourages corporate executives from properly supervising their subordinates. By comparison, the corporate ethos standard directs criminal liability toward only those corporations which are “deserving” of prosecution as demonstrated by their lawless ethos. In this way the corporate ethos standard rewards corporations that police themselves (p. 1157).

The corresponding semantic network looks like the one in Fig. 5.
Fig. 5

Semantic network of example 4 [5, p. 1157]

In this semantic network we see a structure in which distinguishing characteristics of the three operationalisations of the concept of criminal liability of corporations treated in the article are presented (black arguments). The distinguishing characteristics are centred around two focal points: require proof (top right of the network) and corporation (left side of the network). In connection with require proof we have the groups employ miscreant actor (connected to corporation), supervise miscreant actor and commit unlawful act (both connected to high managerial agent). There is an opposition between employ and supervise, which are both connected to the argument miscreant actor, and which are connected to the two operationalisations respondeat superior and MPC standard. The criteria groups are mutually exclusive, as one has a wider scope than the other. Thus, we have here another example of mutually exclusive or opposing knowledge elements connected to the concept under scrutiny.

In connection with the argument corporation we find distinguishing characteristics concerning how the different operationalisations treat the corporations (according to the view presented by the author). Connected to respondeat superior we find arguments indicating that all corporations are treated alike. The same is stated in connection with MPC standard (not distinguish). In opposition to this, corporate ethos standard is characterised by being oriented towards only those corporations that deserve prosecution due to lawless ethos. Again, the characterisations are mutually exclusive, as they define the differences between operationalisations based on different and incompatible criteria.

A third opposition of incompatibilities exists between corporate ethos standard and MPC standard. The first argument is connected to a group of elements stating that the standard rewards a corporation that police themselves, whereas the second argument is connected with a group of elements indicating that the standard discourages corporate executives from supervising subordinates.

In conclusion, we find also in this example a number of instances where the knowledge base of the author must contain not only the knowledge elements that the author herself believes to be the correct and relevant ones when talking about the concept of criminal liability of corporations, but also such knowledge elements, which the author believes to be false. So the legal expert seems to have an individually profiled system of knowledge elements, consisting of believed and non-believed elements.

On the basis of a scrutiny of all four examples it may be stated that as expected on the basis of the theoretical discussion in Sect. 24 we do not find monolithic systems of knowledge elements reflected in the two investigated texts. First, although the two authors agree upon the contours of the concept of criminal liability of corporations, they generally disagree as to the justification of the concept: Becker refutes it altogether, whereas Bucy accepts the idea of having such a concept, but suggests a narrower way of operationalising the concept (thus narrowing the concept as such). So there is a demonstrated lack of identity in the knowledge bases of the two authors concerning the concept of criminal liability of corporations. Secondly, through the way the two authors represent the different knowledge elements in the text they show that they must both hold a variety of knowledge elements connected to the same aspects, but contradicting each other.

In every example, it is easy to see which of the subsets are supported by the author. The data may only be interpreted in the way that the author simultaneously holds knowledge about different possible positions, apart from the one favoured by him- or herself. In the light of the concept of Theory of Mind suggested in Sect. 4 as a cognitive mechanism with relevance for achieving communicative efficiency in a situation, where experts have non-identical knowledge bases this state of affairs is interesting. For it documents that the knowledge bases of the two experts under scrutiny here have the necessary structure and profile for them to function as basis for constructing a relevant Theory of Mind. Our two experts would most likely be able to efficiently process communicative input from experts holding any of the positions presented in their texts, even though the knowledge bases of these other experts would necessarily have a different profiling, as they would have to consider other elements incorrect. The documented structures in combination with the cognitive mechanism thus show how efficient communication may be possible despite lack of identity of the knowledge bases.

Conclusion and Perspectives

In the empirical part of my article I have tried to show that what we find when looking at the individual knowledge applied in academic legal texts is better in accordance with presuming similarity and disagreement than identity of the systems of knowledge elements of the individual legal experts.

These findings indicate that in this kind of communication involving human experts the idea of individual participants in communication functioning in isolation (which has been the basic assumption of traditional terminology) should be abandoned. I am not advocating that we abandon approaches like terminology or other code-oriented approaches altogether, but that we supplement these approaches by others with a different set of basic assumptions. The results of the above scrutiny indicate that we should look more closely at the process of communicative negotiation of meaning also in the field of law. With Thibault [44] one could say that we should be more interested in the conversational dyad of sender and receiver as the relevant locus of meaning than we have traditionally been in the field of LSP studies and in the study of legal communication: specialised knowledge is inherently local and individually profiled, but it is constructed, shaped and reshaped in interaction with the surrounding context [6, p. 26]. A cognitive approach to the study of legal communication should take both of these aspects (the individual’s cognitive basis and the impact of interaction on it) into consideration and take the conditions of the individual cognitive system seriously. Thus, we will have a good chance of describing and explaining academic knowledge and meaning as it really is, with its actual stabilities and dynamics.

One last question to be touched briefly upon here is whether it makes a difference outside the realms of academic description of meaning if we enlarge the focus, as suggested here, to include actual (in)stabilities of meaning due to individuality of expert knowledge. I think it does. Mainly because such an enlargement of focus undermines the possibility of stating that something like a fixed, original meaning laid down in a text and to be extracted from there can at all exist. An approach like the one suggested here (just like other discursively oriented approaches to the description of meaning like the ones by Müller [30] and Christensen [7]) underlines the fact that legal meaning is not basically different from other types of natural language meaning: it is subject to constant cognitive creation and recreation and must therefore be inherently instable. This instability, however, does not lead to discretion in legal meaning. Discursive forces like the group of relevant experts (in casu the judges and lawyers inside a legal system) help restraining dynamic forces and thus help securing the functionability of the system of meanings. Thus individual and social constructions go hand in hand. And by applying approaches and methods like the ones presented here we can hope to better strike the balance between the two, thus preventing ourselves from overemphasising neither the power of the individual nor the power of the social forces in actual meaning construction processes.

Of course, a small analysis like the one presented here may not fulfil this whole task. I have here concentrated upon showing that conflicting aspects of a disputed concept will be simultaneously present in the minds of the participants in actual communication, and that this fact is actually an advantage for achieving communicative precision under the conditions of the human cognitive system. In order to say more about the individual cum social process of construction in human legal communication, the presented individually oriented approach must be supplemented by other approaches like the ones from Social Semiotics mentioned in Sect. 2, but also by other approaches concentrating more upon the social side of meaning constructing in specialised settings. I will here mention only one central candidate, viz. Kuhns ideas of scientific communities as the keepers of paradigms that are discursively refined or discarded of through communication inside and outside of the scientific community adhering to the paradigm [25]. All three phases in the process of scientific progress may be described as constellations of knowledge elements in the minds of individuals participating in the communicative exchanges accompanying and driving the process. And thus the two approaches can definitely inform each other, just like approaches from hermeneutics will be interesting to include when looking more deeply into the process of understanding on the basis of individuals’ Theories of Mind.11 Thus, it is my firm belief that in this case, like always, we can only reach a thorough description of the real-world problems to be solved in actual communication by joining forces across different approaches to the description of individual and social communicative processes.


Roelcke [35, pp. 17–26] talks about approaches focusing upon the influential factors as subscribing to the context model of pragmalinguistics. Studies of this kind typically come from the fields of genre or discourse analysis.


An interesting case of focusing on the communication of the group, but with interest in developments, is the work on detecting new terms and concepts in corpora by looking for words with a specifically high ‘weirdness factor’ [20].


Roelcke [35, pp. 17–26] sees such terminological approaches as adhering to the inventory model of system linguistics.


Concerning the assumption that mental representations do not primarily have representational character, but are always functionally related to situation-based goals of the holder of the representations cf. [22, p. 204, [24, pp. 94–96, 39, p. 234].


For more arguments in favour of memory as a (re-)constructive process, see [34].


Candlin [6, pp. 24–28] states that this means that specialised discourse is characterised by alterity rather than by intersubjectivity.


As a reaction to similar insights in the field of terminology a number of new approaches to terminological description have emerged. See Temmerman [43] as a very substantiated example of a new approach. [43, pp. 22–34] also has an overview over other critical approaches.


Concerning the distinction, see [33, p. 526].


Sowa [40] calls this type of semantic networks assertion oriented networks.


For a deeper discussion of the methodology, see [10].


I thank one of the anonymous reviewers for these suggestions.



The studies underlying this article have been partly undertaken during my visit as a research fellow at Brooklyn Law School, New York in February and March 2007. This visit was made possible by a grant from the Danish foundation Carlsbergfondet. I want to thank Prof. Larry Solan and Rachel Ehrhardt for invaluable help with these studies.

Copyright information

© Springer Science+Business Media B.V. 2009