(University of Maryland, College Park)
1. Introduction
The purpose of this paper is to think about appropriate methodology to be used in information ethics. As a newly growing field of applied ethics, methodological issues in information ethics have many affinities with those in other fields of applied ethics. Wide reflective equilibrium (WRE) is a method recommended in many other fields of applied ethics (such as bioethics) in which we have to adapt ourselves to a new situation created by new technology, and in a recent article, van den Hoven advocated use of WRE in computer ethics (van den Hoven 1997). Even though I agree with him that WRE is a useful method in applied ethics in general and in information ethics in particular, I also think that there is a mis-characterization of the nature of WRE shared by many authors (including van den Hoven). If we sort out ambiguous claims about WRE carefully, I argue, what remains is actually a form of modest foundationalism. This means that the difference between the traditional method of ethics and WRE is not as large as people may assume. I use an example --- the e-mail privacy issue especially in the workplace --- to illustrate my argument.[1]
2. Case issue: e-mail privacy
This kind of methodological discussion tends to become very abstract, and having a concrete issue in mind is helpful in making sense of such abstract materials. The issue I am going to use in my paper is the issue of e-mail privacy.
The privacy issue related to e-mail, especially in the workplace, have been getting increasing attention recently, and an increasing number of papers on this issue are being published from various points of view. Many e-mail users simply assume that the privacy protection of postal mail also applies to e-mail (Cappel 1993; Cappel 1995; Chociey 1997), but the situation is not so simple. As a matter of fact, now many companies have their own e-mail systems to serve their business purposes, and a large part of such companies monitor e-mail, even though not many companies have an explicit monitoring policy (O'Brien 1992; Piller 1993; Hawkins 1997). The main focus of the current debate is the extent of e-mail monitoring to be allowed for those companies.[2]
In the United States, there is a federal law, the Electronic Communication Privacy Act (ECPA) of 1986, which covers e-mail privacy, which leaves a gray zone about e-mail monitoring within a company.[3] There were a few cases in which employees sued the company because of its violation of privacy, but so far the courts have favored monitoring.[4] Some studies suggest that many e-mail users are unaware of this legal situation, and that whether monitoring is legal or not, many people, including managers, consider it unethical (Cappel 1995; Chociey 1997). Why do the courts rule against e-mail privacy? One reason for this is technological differences between e-mail and ordinary mail, which makes the expectation of privacy suspicious.[5]
As many argue, what we need here is a clear legislation or clear company policies (e.g., Glassberg et al 1996).[6] However, this is not just a legal question. We also need moral considerations to decide: what is a morally desirable legislation or policy?; what should we do when there is no law or policy concerning e-mail monitoring?; is monitoring ethical if the company has a monitoring policy?; and so on. The key issue is where to locate e-mail among various ways of communication, like postal mail, phone conversations, and public speech. Simply pointing out similarities and differences between e-mails and other forms of communication is not enough for such considerations. What we need are considerations about why such similarities and differences are morally relevant. There is also the issue of expectations of the parties. Even if users expect e-mail exchanges to be private, that does not mean that the expectation is warranted (though that expectation may be a part of the reason we prefer e-mail privacy). In fact, in the case of a company e-mail system, the company owns the e-mail system and provides its employees with the e-mail service to improve business performance. This may be a good reason why employees should not expect privacy on such an e-mail system. (See the attached table for arguments for and against e-mail monitoring.)
As a ground for assessment, some people advocate Kantian ethics (Thompson et al. 1995), others seek a more consequentialist solution (e.g., Glassberg et al. 1996).[7] Of course, the entire philosophical and legal literature on the privacy issue is also relevant here (Schoman 1984 is a nice anthology of classical papers on this issue). [8]
What should we do to deal with this rather messy body of information? What I am going to offer here is a way to approach this kind of question. However, since the methodological issue related to WRE is the main focus of this paper, I will not attempt to give a definite answer to the e-mail privacy issue (that should be the topic of another paper).
3. Wide reflective equilibrium
To talk about wide reflective equilibrium (WRE), it is convenient to start with the original 'narrow' version introduced by John Rawls (Rawls 1971). Reflective equilibrium is a reflective process between moral principles and considered moral judgments (48). Considered judgments are defined by Rawls as the judgments "rendered under conditions favorable to the exercise of the sense of justice" (47), and persons making such judgments are presumed "to have the ability, the opportunity, and the desire to reach a correct decision (or at least not the desire not to)" (48). We compare the principles and considered judgments, and when we find discrepancies between them, sometimes we change considered judgments, and sometimes we change principles, until "our principles and judgments coincide" (20).
This 'narrow' version of reflective equilibrium has been criticized as too subjective to reach a reliable moral judgment (Hare 1973; Singer 1974), and Norman Daniels's WRE was introduced to remedy the problem (Daniels 1980).[9] WRE, according to Daniels, is an interplay among three components (85-86): (a) considered moral judgments, (b) a set of moral principles, and (c) a set of relevant background theories. Thus, when we find some discrepancy between (a) and (b), we appeal to background theories in (c) to decide what to revise. Among background theories, Daniels includes general social theory, theory of moral development, theory of the role of morality in society, theory of persons, theory of procedural justice, and so on (88). [10]
Let us rephrase these three components in terms of the e-mail privacy issue. Considered judgments are judgments we make on each concrete ethical question, like "if a company has an explicit e-mail monitoring policy and employees are aware of that policy, is monitoring ethically admissible?" (the question can be more specific if necessary). Moral principles relevant here include the protection of privacy,[11] the Kantian concept of autonomy, the principle of utility, and so on. These principles vary in terms of the scope and the level of abstraction, and this is an issue we need to deal with later. Finally, the background theories to be considered include sociological and psychological theories about the role of privacy in our social and personal life, and other general theories about how society works.
What should we do to reach an equilibrium among these components, exactly speaking? Here, WRE becomes very ambiguous, and this is where some critics attack it. The main problem here is in the ambiguity about the source of justification in WRE. Let us look at the issue in detail.
4. WRE and the sources of justification
According to Daniels, there are two roles played by WRE (Daniels 1980, 101-103). One is "analytic and explicative" (101), namely a clarification of the structure of moral beliefs. For example, WRE can reveal unnoticed background theories which influence our considered judgments. When we realize the existence of such background theories, we may want to change the considered judgments. To use the e-mail privacy case, a person who is opposed to a private use of an in-company e-mail system may find out through WRE that her opposition is based on an assumption that such an e-mail use reduce productivity. By making this assumption conscious, she can actually investigate the issue empirically, and if she does not find a correlation between productivity and private use of e-mail, now she has a reason to doubt her original considered judgment. I have no quarrel with this explicative role of WRE.
The other role of WRE, according to Daniels, is to serve as "a basis for a coherence account of moral justification" (Daniels 1980, 102). Coherence is "evidential" for the correctness of the view arrived at by WRE (103). Here, Daniels is appealing to a common distinction between 'foundationalism' and 'coherentism' about the source of moral justification. In a typical distinction, foundationalism is a view that the source of justification of a moral judgment is some self-evident and irrevisable foundation on which all moral judgments should be based. On the other hand, coherentism is characterized as a view that the justification comes from coherence among various components of our system of moral (and non-moral) judgments. Foundationalism has been attacked because there is no such self-evident foundation everyone can be satisfied with. With this context in mind, Daniels seems to be claiming that WRE is a coherentist model of justification which avoids the problem of foundationalism. Van den Hoven clearly takes this view when he applies WRE to computer ethics, saying that there is no foundation and "there are only relations of noncontradiction, consonance and connectedness" (van den Hoven 1997, 243). However, such a simple coherentist model of justification does not seem promising as a method to solve practical moral problems.
A charge frequently raised against the coherentist justification model of WRE is that it is useless in choosing among alternatives. We can think of cases in which different sets of initial considered judgments may result in different WRE, and even when we agree upon the set of initial considered judgments to start with, we still may end up with different WRE, if the only source of justification is the coherence of the attained equilibrium (Little 1984; Haslett 1987).[12] Think about two people whose initial considered judgments about the e-mail privacy are totally different; such as a person who thinks that e-mail monitoring is always ethical even when there are legislations and policies against monitoring, and another person who thinks monitoring is always unethical even when legislations and policies explicitly permit it. By taking moral principles and background theories into account, they may relax their extreme positions.[13] However, the resulting equilibrium point may still be different (the latter person is likely to end up with admitting more privacy than the former person). Even when two people start from the same set of considered judgments about e-mail privacy, when they come to know Kantian ethics, it is possible that one finds it very plausible and shift toward more protection of privacy, and the other finds it totally unconvincing and simply dismisses the principle.
Both Rawls and Daniels were disturbed by the problem with different sets of initial considered judgments, and propose that WRE should be used where there is a substantial consensus (Rawls 1980, 519; Daniels 1979b, 281). Given the widespread lack of consensus across different cultures on moral issues, the consensus they are talking about seems to be a consensus in the mainstream Western culture. Such an appeal to `our' consensus may elicit a charge of ethnocentrism, as Nielsen worries (Nielsen 1977; 1982).
Nielsen prefers to appeal to shared non-moral background theories (Nielsen 1982; 1993). He admits that where there is an incommensurability in world views, there is no hope of attaining a WRE (Nielsen 1993, 328). However, where world views are commensurable, we can use this as a fairly fixed point to attain an agreement in moral issues. Such an appeal to shared non-moral background theories may not eliminate all disagreements, but then we simply pour more and more information and arguments into the discussion until an agreement is reached.
Though interesting and suggestive, Nielsen's argument seems far from successful in replying to the criticisms it intends to answer. The usefulness of non-moral background theories in resolving fundamental moral disagreements is, to say the least, very questionable. Holmgren analyzes various cases in which a factual agreement can be used to resolve a moral disagreement, and concludes that in none of these cases can a factual agreement alone do the job; we need a substantial moral agreement on certain matters, too (Holmgren 1989, 51-54). To illustrate, in the case of e-mail privacy, suppose that a psychological theory says that e-mail privacy is very important for employees' mental health. This factual claim has no implication about how to revise our moral principles or considered judgments unless we also maintain a normative belief like "it is desirable that employees are mentally healthy". Of course, may of us would agree with this particular normative judgment, but the point is that there is nothing incoherent in rejecting this normative judgment and therefore in dismissing the psychological claim about mental health as irrelevant.
We need something more than a mere coherence to evaluate various equilibria. Here, I agree with several critics of WRE that WRE is consistent with a form of foundationalism, namely modest foundationalism (DePaul 1986; Ebertz 1993).[14] Unlike classical foundationalism, modest foundationalism does not assume that the foundation is irrevisable or self-evident. Rather, the foundation justifies other beliefs and judgments because of the relative reliability of the foundation. Actually, few present-day foundationalists see their foundations irrevisable, thus we can regard them as modest foundationalists, too. For example, foundationalists like Hare (1972) and Brandt (1990) admit that our considered judgments play an important heuristic role in refining basic principles. All these considerations suggest that the difference between classical moral methodology and WRE is not as large as it seems.
5. A modest foundationalist model of moral reasoning
Given the discussion in the previous section, let me describe a reasoning process which seems to me to be a better description of WRE. I call this 'a modest foundationalist model' of moral reasoning. Even though it is applicable both to an individual decision process and to a collective decision process, I concentrate here on a collective decision process in which we look for standards we can endorse collectively.
In this model, we start from comparing everyone's considered judgments. In a novel and controversial issue like e-mail privacy, we find a variety of different views on certain concrete cases, though we may also find a certain level of agreement on other cases.[15] Next thing to do is to ask each of us the reason why the person judged in the way he/she did. For example, a person who supports e-mail monitoring within a company may give a reason, "because the system is owned by the company." Here, the person is implicitly appealing to a principle that a company can look into the content of what it owns. Since the principle is invoked as a reason for the particular judgment, here is a clear direction in the justificatory relationship: the principle is justifying the particular judgment by giving a reason. When asked why a company can look into the content of what it owns, the person may appeal to even higher principle of the property right, and again the higher-level principle is giving the reason why the lower-level principle should be accepted. The point here is that invoking a higher level principle in answering a practical question does not make much sense unless the principle is conceived as reason-giving. Here is a justificatory asymmetry between principles and considered judgments, which makes the model foundationalist.
However, such principles invoked as reasons may be challenged in various ways. First, the invoked principle may be inconsistent with some of the considered judgments the person does not want to change. This is a sign that something went wrong in the search for a reason: that is, the person invoked a principle which he/she did not mean to invoke. For example, we may ask the person mentioned above, "then what about a letter written on a letterhead of the company? May the company read it?" If the person feel uncomfortable to answer yes, that suggests that he/she needs a more specific reason which can differentiate an e-mail system from a letterhead.[16] Second, background theories come in here. That is, the invoked principle may be base on a suspicious background theory (we saw an example of this already). Also, philosophical considerations about relative advantages and disadvantages of alternative principles are relevant here (Rawls 1974/1975, 8). This process of refinement is 'explicative' (rather than justificatory) in Daniels's sense, because this is a search for (and a refinement of) our implicit principles and background theories to be examined.
In the case of collective decision-making, there is a further constraint on the kind of reasons and principles to be invoked: namely, the principles should be such that all parties can collectively endorse them. Thus, we cannot appeal to self-interests or idiosyncratic ideals as reasons, unless these interests and ideals are balanced with those of others. For example, if a person insists that the right of privacy is absolute and should be protected at any cost, while other people do not find the ideal convincing, we can never reach a collective endorsement. What we need here is a method to deal with different interests and ideals in a way everyone can agree with. This method may be a certain decision process (such as a democratic decision-making) or an established rule of comparing interests and ideals (such as preference utilitarianism).[17]
The result of adding these structures and constraints on WRE is that we can narrow down the range of resulting equilibria. The source of justification here is the relative acceptability (in the sense of collective endorsement) of the basic principle(s) arrived at in this way.[18] I do not think that the process can single out a unique equilibria as the best alternative, but if coherence is the only source of justification, we would end up with a much worse situation.
6. Concluding remarks
In this paper, I first described WRE as introduced by Daniels, with his three components of WRE. Daniels distinguished two roles of WRE, namely an explication and a coherentist justification. The latter role has been criticized because such a view of justification makes a choice among alternatives very hard. I suggested that WRE can be better understood as a form of modest foundationalism, and I also described a model of moral reasoning based on my suggestion.
A final remark needs to be made about the implication of this method to the e-mail privacy issue, though, as I said earlier, the purpose of this paper is not to offer a solution to the issue. I think most of us value privacy in some sense, but at the same time most of us hesitate to say that privacy is an absolute right.[19] Some balancing seems to be needed here. Appeals to ideals like Kantian notion of autonomy is at most a partial answer to the situation. Such ideals should be respected, but in a way in which they are balanced with other interests and ideals. It seems to me that there is no simple answer here, and that is the point of using a complicated procedure like WRE to arrive at what we can collectively endorse.
Notes
[1] As this introduction makes clear, even though I deliberately chose an issue in information ethics as my example, basically the same methodological argument applies to any kind of applied ethics. The development of information technology certainly has created a set of new interesting ethical issues, but WRE is (partly) a method to extend existing ethical systems to a new situation, and we do not have to re-create the tool of extension itself.
[2] This does not mean that the issue is irrelevant in the academic sphere. As Veeder (1995) points out, what we eventually need is a coherent policy which covers the academic, governmental and private sectors.
[3] Even though ECPA protects the privacy of e-mail in general, there are a few important exceptions permitted in the law: monitoring is allowed (1) if a provider needs to monitor e-mail to make the service possible or to protect the property of the provider; (2) if the monitoring is an extension of the ordinary course of business; or (3) if there is a prior consent of one party to the communication. See Rodoriguez 1998, 1450-1460 for a detailed review of the content of this law.
[4] See, for example, Sipior and Ward 1995 and Pivec and Brinkerhoff 1999 for several such cases. Most of these cases are based on the common law tort of invasion of privacy, not on ECPA. This common law requires a reasonable expectation of privacy, and in most cases the court find such an sense, but at the same timexpectation regarding to e-mail is unwarranted.
[5] Such differences include: (1) e-mail needs a service provider who maintains the server; (2) unlike ordinary mail, content of e-mails is backed up routinely; and (3) it is relatively easy to monitor e-mail, and sometimes such monitoring becomes necessary for maintenance. Weisband and Reinig 1995 lists other features of e-mail which justify the expectation of privacy (such as the use of password and certain userinterface which emphasize the analogy between e-mail and postal mail).
[6] In the United States, a bill called the Privacy for Consumers and Workers Act (PCWA) is proposed at the federal level, which intends to protect employees' privacy more effectively. However, this bill has not gained enough supporters so far. See Rodoriguez 1998, 1464-1465.
[7] I could not find a genuine utilitarian approach to the e-mail privacy issue in the literature.
[8] There is also a recent proposal that Habermas's notion of the public sphere is useful for analyzing this issue (Primeaux 1998). Even though it is an interesting attempt, what Primeaux adds by introducing the notion of public sphere seems little more than the right of privacy in the sense of control of personal information.
[9] Daniels notes that Rawls himself introduced the distinction between narrow and wide reflective equilibria (Daniels 1979a, 257 n.2), but Rawls's version seems to have a slightly different emphasis. See Rawls 1974/1975, 8.
[10] As you may notice, some of these theories are moral theories with normative content. Daniels requires 'independence' of these theories in the sense that these background theories are not based on considered judgments in (a) (otherwise appeal to such background theories would be of little help). This is called the 'independence constraint' (Daniels 1080, 87).
[11] As many authors point out (e.g. Introna 1997), the notion of privacy has been defined in many different ways in the literature, and there is no obvious way to pick up one definition rather than another. In WRE, what we may want to do is to list all different definitions as different principles. Thus, 'protection from being watched' and 'protection of the right of control over self-information' may be different entries in (b).
[12] In applied ethics, it is convenient to distinguish two different types of ethical problems:(a) Individual decision-making: problems of individual decision-making in a hard case; (b) Collective decision-making: problems of public policy-making when there are conflicts of interest and/or ideals. Of course, these two types of problems are interrelated (e.g., if there is a definite public policy, it is easy to make an individual decision), but there is an important difference between the goals of these two types. In (a), the goal is to find a solution to a hard case which you can commit yourself to. In (b), the goal is to find a solution which all affected parties can collectively commit themselves to. WRE can be used in both cases, in different ways. Little's criticism is mainly directed toward the context of collective decision making, but a related problem occurs when an individual decision-maker recognizes alternative ways to attain an equilibrium.
[13] Please note that this is not necessarily the case if we adopt the simple coherentist view of WRE. A person with extreme views may attain an equilibrium by rejecting any moral principle inconsistent with his view.
[14] Ebertz regard considered judgments, rather than principle, as his foundation, so his version is different from the version I advocate in the next section. I do not accept Ebertz's version because it seems to be falling back to the subjectivism Hare (1973) and Singer (1974) criticized.
[15] If we are in an agreement on every concrete issue, namely in all considered judgments, then there is not much point in asking ethical questions in philosophical terms, even though such considerations may lead us to a better understanding of the situation, or to a revision of our present consensus. For example, in the e-mail privacy issue, no one in the debate really wants to challenge the privacy protection of postal mail.
[16] I owe this example to Thompson et al. 1995, 161-162. Of course, if the person cannot find any good reason to differentiate these cases, the person may have to change the original considered judgment about the letterhead case. In this sense, even the firmest conviction is revisable.
[17] Hare's notion of universalizability (1963) and Rawls's notion of original position (1971) are proposed to serve roughly the same purpose, but I am not happy with either of them. Hare's universalizability is not really a collective decision-making in the sense that a fanatic can stick to his idiosyncratic ideal, if his ideal is firm enough. The specificities of Rawls's original position are determined in a reflective equilibrium, and as Rawls himself admits (Rawls 1980), his reflective equilibrium is based on particular political ideals. Details of these criticisms are way beyond the scope of the present paper.
[18] Note that I am not appealing to initial plausibility of considered judgments. Considered judgments are used to find a more acceptable principle which presumably we implicitly adopt.
[19] Here, another concern is cultural/national differences on privacy issue. See Milberg et al. 1995 for a recent study on nationality and privacy concern. In the case of e-mail privacy, a culturally relative solution is not satisfactory. E-mail, along with other internet resources, makes a cross-cultural interaction much easier than before, and what we need here is a standard which people from different cultures can collectively endorse.
References
Brandt, R.B. (1979) A Theory of the Good and Right. Oxford: Clarendon Press.
--. (1990) "The science of man and wide reflective equilibrium" in Ethics 100, 259-278.
Cappel, J.J. (1993) "Closing the e-mail privacy gap" in The Journal of Systems Management 44 no.12, 6-11.
--. (1995) "A study of individuals' ethical beliefs and perceptions of electronic mail privacy" in Journal of Business Ethics 14, 819-827.
Chociey, P.A. (1997) "'Who's reading my e-mail?': a study of professionals' e-mail usage and privacy perceptions in the workplace" in IEEE Transactions of Professional Communication 40, 34-40.
Daniels, N. (1979a) "Wide reflective equilibrium and theory acceptance in ethics" Journal of Philosophy 76, 256-282.
--. (1979b) "Moral theory and the plasticity of persons" in The Monist 62, 265-287.
--. (1980) "Reflective equilibrium and Archimedean Points" in Canadian Journal of Philosophy 10, 83-103.
Daniels, N. ed. (1975) Reading Rawls. New York: Basic.
DePaul, M. R. (1986) "Reflective equilibrium and foundationalism, in American Philosophical Quarterly 23, 59-69.
Ebertz, R. (1993) "Is reflective equilibrium a coherentist model?" in Canadian Journal of Philosophy 23, 193-214.
Glassberg, B.C., Kettinger, W.J. and Logan J.E. (1996) "Electronic communication: an ounce of policy is worth a pound of cure" in Business Horizons 39 no. 4, 74-80.
Hare, R.M. (1963) Freedom and Reason. New York: Oxford University Press.
--. (1972) "The argument from received opinion" in Essays on Philosophical Method. Berkeley, University of California Press.
--. (1973) "Rawls' theory of justice", in The Philosophical Quarterly 23, reprinted in Daniels 1975, 81-107.
Haslett, D.W. (1987) "What is wrong with reflective equilibrium?" in The Philosophical Quarterly 37, 305-311.
Hawkins, D. (1997) "Who's watching now?" in US News and World Report Sept. 15, 56-58.
Holmgren, M. (1989) "The wide and narrow of reflective equilibrium", in Canadian Journal of Philosophy 19, 43-60.
Little, D. (1984) "Reflective equilibrium and justification", in Southern Journal of Philosophy 22, 373-387.
Milberg, S.J., Burke, S.J., Smith H.J. and Kallman, E.A. (1995) "Values, personal information privacy, and regulatory approaches" in Communications of the ACM vol. 38 no. 12, 65-74.
Nielsen, K. (1977) "Our considered judgments", in Ratio 19, 39-46.
--. (1982) "Considered judgments again", in Human Studies 8, 109-118.
--. (1993) "Relativism and wide reflective equilibrium", in The Monist 76, 316-332.
O'Brien, T. (1992) "E-mail to be foundation of new information highways" in Network World vol. 9 no.21, 23-24.
Piller, C. (1993) "Bosses with x-ray eyes" in Macworld vol. 10 no. 7, 118-123.
Pivec, M.E. and Brinkerhoff, S. (1999) "Email in the workplace: limitations on privacy" in Human Rights 26, 22-25.
Primeaux, D. (1998) "Using an alternative ethical paradigm for analysis; an example regarding e-mail privacy issue" in Computers and Society vol. 28 no.2, 52-55.
Rawls, J. (1971) A Theory of Justice. Cambridge: Harvard University Press.
--. (1974/75) "The independence of moral theory", in Proceedings and Addresses of the American Philosophical Association 48, 5-22.
--. (1980) "Kantian constructivism in moral theory: The Dewey Lectures 1980", in The Journal of Philosophy 77, 515-572.
Rodriguez, A.I. (1998) "All bark, no byte: employee e-mail privacy rights in the private sector workplace" in Emory Law Journal 47, 1439-1473.
Schoeman, F. D. ed. (1984) Philosophical Dimensions of Privacy: An Anthology. Cambridge: Cambridge University Press.
Singer, P. (1974) "Sidgwick and reflective equilibrium" in The Monist 58, 490-517.
Sipior, J. C. and Ward, B.T. (1995) "The ethical and legal quandary of email privacy" in Communications of the ACM vol. 38 no.12, 48-54.
Thompson, J.A., DeTienne, K.B. and Smart K.L. (1995) "Privacy, e-mail, and information policy: where ethics meets reality" in IEEE Transactions of Professional Communication 38, 158-164.
van den Hoven, J. (1997) "Computer ethics and moral methodology" in Metaphilosophy 28, 234-248.
Veeder, S.B. (1995) "Electronic mail and privacy" in The Journal of Academic Librarianship 21, 123-126.
Weisband, S.P. and Reinig B.A. (1995) "Managing user perceptions of email privacy" in Communications of the ACM vol. 38 no. 12, 40-47.
Arguments for e-mail monitoring
1. E-mail system is owned by a company, so the company can use it as it wants.
2. E-mail system is supposed to be used for business purposes. It is legitimate to monitor e-mail to make sure that it is not used against that purpose (to waste working time, to leak company secret, etc.)
3. E-mail is so easy to monitor, compared with other means of communication. It is simply wrong to expect e-mail privacy.
4. Analogy between e-mail and postal mail is limited. E-mail in the workplace is more like speaking in public (or at least speaking in front of your boss).
5. Having an explicit policy is against the purpose of monitoring, because the purpose is to catch wrong-doings of employees.
6. When some employee uses e-mail for illegal purposes (like sexual harassment), the company may held to be responsible.
Arguments against e-mail monitoring
1. Privacy is a basic human right, violation of which means a loss of dignity.
2. Monitoring has negative effects on employees' productivity; it reduces their creativity, responsibility, etc. Monitoring limits free flow of idea in a company. Employees may even work harder by being trusted.
3. Ease of monitoring does not make monitoring ethical (ease of killing does not make killing ethical).
4. Given employees' expectation of e-mail privacy, monitoring is a kind of deception.
5. Many companies do not have monitoring policy. It is unethical to monitor e-mail at least without an explicit policy.
6. At least the monitoring should be restricted to those whose behaviors are suspicious.
7. Privacy is an important part of our personal identity. We need privacy for our mental health.
8. Privacy is an important part of our (modern) social life. We manage relationships with others by controlling our personal information. Monitoring destroys this basic feature of our social life.