What Is It to Know Someone? David Lauer, Freie Universität Berlin -berlin.de

What Is It to Know Someone?
David Lauer, Freie Universität Berlin
[email protected]
March 2014
1. At the very beginning of his Stanford Encyclopedia of Philosophy entry on „Knowledge
How“1, Jeremy Fantl writes that it is “common in epistemology to distinguish among three
kinds of knowledge”, which he characterizes as follows:
“There’s the kind of knowledge you have when it is truly said of you that you
know how to do something—say, ride a bicycle. There’s the kind of knowledge
you have when it is truly said of you that you know a person—say, your best
friend. And there’s the kind of knowledge you have when it is truly said of you
that you know that some fact is true—say, that the Red Sox won the 2004 World
The next sentence of the entry reads: “Here we will be concerned with the first and last of
these kinds.” This casual passing over the second of the three kinds of knowledge strikes me
as somehow characteristic of current epistemology. Surely claims to know others—or the denials of such claims—play an important role in our lives, and it should be a worthy philosophical project to ask what we mean by saying things like “She knows me really well”, “I hardly
know you”, or “You don’t know me at all”. The question I would like to address, therefore, is
the one the title of my paper states: What is it to know someone?
2. I should point out from the start, however, that the paper is going to take but a first, very
elementary step in the direction of a general answer to its question. As the examples I just
gave show, knowledge of others2 is gradable. The question “Do you know her?” admits of
answers ranging on a spectrum from “inside and out” to “hardly” and “not at all.” Although
such methodological commitments are never innocent in philosophy, I shall start from the
lower end of the spectrum. So the task of this paper is to give an account of what it minimally
Jeremy Fantl: “Knowledge How”, in: Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy. URL =
http://plato.stanford.edu/entries/knowledge-how/. First published 12/04/2012, retrieved 03/23/2014.
I will speak of “knowledge of others” interchangeably with “knowledge of (other) persons”.
is to know someone, of what it takes for something to be a case of somebody knowing someone überhaupt. I suggest to understand the richer and more demanding cases—e. g., what
characterizes the kind knowledge you have of your best friend, to use Fantl’s telling and altogether non-coincidental example—by building on this minimal account, but doing this is a
project that goes well beyond what can be achieved within the confines of a single paper, and
thus will have to wait for other occasions.
3. I will argue for the following claims, which are at the same time the titles of the paper’s
main sections: (I) Knowledge of others seems to be a form of knowledge of objects of a distinctive kind, i.e., other persons. So we should establish what the fundamental form of
knowledge of objects of this kind is. (II) The fundamental form of knowledge of oneself as a
person—self-knowledge, knowledge of oneself as “I” —is knowledge from spontaneity, not
from receptivity. We will argue for the thesis that this is true of the fundamental form of
knowledge of persons in general, hence also of the fundamental form of knowledge of others.
(III) To make this plausible, we will show that the everyday conception of what it is to know
someone “personally” determines this form of knowledge as knowledge by interaction, and
distinguishes it from knowledge by acquaintance. (IV) And indeed, our considerations show
that the fundamental form of knowledge of others is knowledge by interaction, which is
knowledge from joint spontaneity within a second-personal nexus of two persons performing
one act together. (V) The conceptual representation that grounds this fundamental form of
knowledge of another person is the second-person pronoun, “you”. So the fundamental form
of knowing someone is knowledge of her as as “you”, second person knowledge. But it turns
out that second person knowledge is essentially knowledge of the same form as knowledge of
oneself as “I”, i. e. first person knowledge. (VI) Thus, in knowing each other as “you” and
“I”, two persons united in a second-personal nexus share in one act of grasping the same
thought from two perspectives. Thus it turns out that what it is to know someone, in the fundamental sense, is not a case of someone standing in a relation to an object at all. Rather, if A
and B are united in a second-personal nexus of a joint act, they mutually know each other as
the joint subject of this act. I will conclude with some remarks about the possible importance
of this finding for the philosophical debate on knowledge of other minds.
I. Knowledge of others seems to be a form of knowledge of objects of a distinctive kind, i.e.,
other persons
4. It is customary in epistemology to distinguish between kinds of knowledge on the basis of
differing logical forms of knowledge ascriptions. Here are some examples of such ascriptions
and the forms of knowledge they attribute to a subject S: “S knows that the Red Sox won the
World Series” (knowledge-that), “S knows how to ride a bicycle” (knowledge-how), “S
knows what it’s like to have migraine” (knowledge what it’s like), “S knows where to find the
best mushrooms in this forest” (knowledge-wh). Debates rage wildly in epistemology about
how these respective kinds of knowledge are to be understood, whether a unified account of
them is available, and especially whether it is possible to understand some of these kinds as
being reducible to one of the others.3 We can put these debates aside for the purposes of this
paper. What is crucial for our topic is to note that the kind of knowledge we are interested in,
knowledge of others, is attributed by locutions of still another form, namely “S knows O”.
Such attributions characterize the knower (S) as standing in the relation of knowing to an object (O). This kind of knowledge is therefore often referred to as knowledge of objects. (The
notion of an object applicable here is of course purely formal.)
5. As a general characterization of the kind, this seems fine. However, it is way too general in
order to tell us much about what it is to know someone. For there appear to be different species of knowledge of objects, depending on the kind of object known. One of the forms
knowledge of objects may take, for example, is the one Bertrand Russell famously termed
“knowledge by acquaintance”:
„I say that I am acquainted with an object when I have a direct cognitive relation
to that object, i.e. when I am directly aware of the object itself. When I speak of
a cognitive relation here, I do not mean the sort of relation which constitutes
judgment, but the sort which constitutes presentation. In fact, I think the relation
of subject and object which I call acquaintance is simply the converse of the relation of object and subject which constitutes presentation. That is, to say that S
Examples of such debates include, of course, the huge literature on the concept of “knowledge what it’s like”
sparked by Frank Jackson’s knowledge argument (cf. Frank Jackson, “What mary didn't know”, in: The Journal
of Philosophy 83 (1986), pp. 291-295). Just as prominent in recent years is the debate on whether knowledgehow is reducible to knowledge-that (with Jason Stanley arguing in favour of this thesis, cf. his Know How, Oxford: Oxford University Press 2011, ch. 2, and Stephen Hetherington arguing against it, cf. his How to Know: A
Practicalist Conception of Knowledge, Oxford: Blackwell 2011. The debate on the concept of knowledge-wh
currently seems to pick up momentum, cf. Jonathan Schaffer, “Knowing the Answer”, in: Philosophy and Phenomenological Research 75 (2007), pp. 383–403; Berit Brogaard, “What Mary Did Yesterday: Reflections on
Knowledge-wh”, in: Philosophy and Phenomenological Research 78 (2009), pp. 439-467.
has acquaintance with O is essentially the same thing as to say that O is presented to S.“4
Michael Tye aptly sums up this classic statement of Russell’s by saying that I am acquainted
with something if and only if “I have […] encountered [that] thing in experience”.5 I take it
that in order for acquaintance to count as a kind of knowledge, “presented to S” or “encountered by S” here must imply more than that S’s sense organs have been causally affected by O,
where this would not necessarily entail S’s being aware of the affection. Rather, O must figure
in S’s awareness, it must be distinguishably present in her conscious perceptual experience of
her surroundings. If one is convinced—like me—by Wilfrid Sellars’s claim that in mature
human beings, all conscious awareness, not just of universals, but even of particulars, is (in
some sense) a conceptual affair,6 one can say that S is acquainted with O if and only if O is
perceptually present to S in such a way as to afford S with a perception-based conceptual capacity to refer to O demonstratively as “this-such here”, enabling her to keep track of O
through space and time.7 Thus, S is acquainted with O if and only if S has the ability to form
conceptual de re mental states about O. The form of knowledge sustained by this form of reference is to know the object in question through standing in a perceptual relationship to it
which enables one to demonstratively locate it in egocentric space.
6. This conceptualist construal of the concept of knowledge by acquaintance does not assimilate this form of knowledge to knowledge-that. Even Russell maintains that having
knowledge by acquaintance of an object will, under normal circumstances, go hand in hand
with acquiring knowledge-that about the object. Simply by being visually presented with a red
tomato on my breakfast plate, I will normally know that there is a tomato, that it is red, that it
is lying on my plate, and so on. And yet, Russell contends, this does not always have to be the
case: “Knowledge of things, when it is of the kind we call knowledge by acquaintance, is es-
Bertrand Russell: “Knowledge by Acquaintance and Knowledge by Description”, in his Mysticism and Logic
and Other Essays, London: Allen & Unwin 1917, pp. 209-232, here pp. 209 sq. First published in Proceedings
of the Aristotelian Society 11 (1910), pp. 108-128, here pp. 108 sq.
Michael Tye, Consciousness Revisited. Materialism Without Phenomenal Concepts, Cambridge (Mass.): MIT
Press 2009, p. 101.
Cf. Wilfrid Sellars, “Empiricism and the Philosophy of Mind”, in his Science, Perception, and Reality, Atas-
cadero: Ridgeview 1991, pp. 127-196, § 29. In fact Sellars says that it is “a linguistic affair”, which, as far as
Sellars’s claim is concerned, amounts to the same thing.
Cf. Gareth Evans, The Varieties of Reference, Oxford: Oxford University Press 1982, chap. 6. Compare Tye,
Consciousness Revisited, p. 100.
sentially simpler than any knowledge of truths”.8 And in a sense that is right, for even if we
claim that the form of my conscious experience of the tomato must be conceptually shaped,
presenting the tomato to me, for example, as being red and lying on my plate, we can still
insist that my experience makes these facts available to my faculty of judgment without as
such constituting an exercise of that faculty. My acquaintance with the red tomato lying on
my plate provides entitlement for and thus grounds my true judgment or my belief that there
is a red tomato lying on my plate, without as such being a judgment or a belief itself.9
7. Russell himself is clear that acquaintance is just one particular species of knowledge of
objects: knowledge sustained by having encountered the object in one’s conscious perceptual
experience. Only a certain kind of object offers itself to be known in this particular way, objects that are as much as capable of presenting themselves in one’s perceived surroundings
(paradigmatically, material particulars).10 For other kinds of objects however, acquaintance is
neither sufficient nor necessary in order for them to be known, nor in fact even possible. It is
not sufficient, for example, in the case of knowledge of places or spatial relations between
places. If I say that I know the way to San José, this seems to imply, on any ordinary understanding of these words, not just that I have encountered it in my perceptual experience, but
that I know how to get to San José (a case of knowledge-how). Likewise, if I claim to know
Paris, this would (I take it) normally be understood to imply, not just that I have encountered
the city in my perceptual experience, but that I have some kind of practical grasp of the city’s
Bertrand Russell, The Problems of Philosophy, New York: Henry Holt 1912, p. 72.
Cf. John McDowell, “Avoiding the Myth of the Given”, in his Having the World in View. Essays on Kant,
Hegel, and Sellars, Cambridge (Mass.), London: Harvard University Press 2009, pp. 256-272, for a conception
with this general shape. Another thing Russell also maintains, namely that knowledge by acquaintance is “logically independent of knowledge of truths” (Russell, Problems of Philosophy, p. 72) cannot be upheld on the
conceptualist construal of acquaintance sketched here, at least not as a general thesis. My being acquainted with
an object O may be logically independent of knowledge of any particular truth about O, but not of my general
capacity to acquire knowledge of truths, since demonstrative conceptual capacities of the kind that sustain acquaintance cannot be understood independently of being employed in acts of the faculty of judgment, which is
the faculty of acquriring non-accidentally true beliefs, i. e. knowledge-that.
As is well known, Russell held that the only things we can ever be acquainted with are sense data. I will ignore
this commitment of Russell’s and his putative reasons for it (for an argument how this can be justified, see John
McDowell, “Singular Thought and the Extent of Inner Space”, in his Meaning, Knowledge, and Reality, Cambridge (Mass.), London: Harvard University Press 1998, pp. 228-259).
geography so as to know my way around it.11 This just is what knowing objects of this kind is.
Then there are certain types of objects that—unlike places—I cannot be acquainted with at all,
because they are not the kind of mind-independent thing that one can stand in a perceptual
relation to. Pain states, sensations, or feelings, for example, are perfectly good objects in a
formal sense: we can refer to them, making them logical subjects of various thoughts we entertain about them. And of course, qua objects, they can be known. But, as Wittgenstein
taught us, under no circumstances should we let ourselves be tempted into conceiving of this
form of knowledge as a case of acquaintance, as being sustained by my standing, in exercising
a quasi-perceptual faculty of “inner sense”, in a perceptual relation to a concrete particular – a
particular which then has to be pictured as an object “within” me, strangely lacking the mindindependence that characterizes objects of “outer sense”.12 Rather, to claim that I know, say,
the pain of defeat or the remorse that goes along with a bad hangover, is to claim that I have
undergone, that I have lived through these emotions. Again, this just is what knowing objects
of this kind is. Further kinds of objects knowledge of which cannot consist in being acquainted with them include such objects as can only be known—as Russell termed it—by description, that is through knowledge-that. Thus, if I say that I know the history of the First Punic
War, this obviously means that I know a great many historical truths about this war (a case of
knowledge-that). And that seems to be the only way in which objects like the histories of
events in the distant past can ever be said to be known. It would be a category mistake to try
to conceive of knowledge of this kind of object on the model of acquaintance, or lived experience, or some kind of practical grasp.
8. We can summarize our initial considerations by adapting a formulation by Gareth Evans
and say that for every kind of object, there is a fundamental form of knowledge of objects of
that kind: perceptual acquaintance in the case of material particulars, lived experience in the
Especially concerning the former example it could be argued that acquaintance is not even necessary for
knowing this kind of object, for I may have acquired knowledge of how to get to San José without ever encountering the way in my experience, by using maps or models, for example.
Cf., for example, Ludwig Wittgenstein, Philosophische Untersuchungen / Philosophical Investigations, 3rd
edition, Oxford: Blackwell 2006, § 293. To be sure, Wittgenstein sometimes seems to be saying something much
stronger than this: that sensations and the like are no objects and cannot be known at all. Cf. John McDowell,
Mind and World, 2nd edition, Cambridge (Mass.), London: Harvard University Press 1998, pp. 36-38, for a
sketch of how to read these passages in a way that makes them compatible with the present approach; and David
H. Finkelstein, Expression and the Inner, Cambridge (Mass.), London: Harvard University Press 2003, for a full
case of sensations, and so on. The fundamental form of knowledge of objects of a certain kind
is that form of knowledge on the basis of which an object of that kind is individuated, that is,
identified as an object of the particular kind and distinguished from all other objects of its
kind.13 Now, as I said, knowledge of others seems to be a form of knowledge of objects of a
distinctive kind, i.e., of other persons. So in order to make headway with our question about
what it is to know someone, we must ask: what is the fundamental form of knowledge of other persons?
II. The fundamental form of knowledge of oneself as a person—self-knowledge, knowledge
of oneself as “I”—is knowledge from spontaneity
9. Having rephrased our question in this way, we are now in a position to see that there is a
debate in contemporary philosophy to which we can turn for clarification. For if what we are
looking for is an account of the fundamental form of knowledge of other persons, we are
looking for a species of the fundamental form of knowledge of persons in general. And there
is of course a debate on the fundamental form of knowledge that every human being has of at
least one particular person, namely, herself. I will recapitulate what I take to be some of the
fundamental philosophical insights concerning the fundamental form of knowledge of persons
that is self-knowledge. The ideas I am going to draw on are not uncontroversial. They represent just one of several approaches advocated in the current debate (roughly, it is the approach
that Brie Gertler, in her useful survey of the field, calls the “rationalist” conception of selfknowledge).14 However, it is the approach I favor, and in the present context, I will simply
take it for granted. My aim here is not to discuss and defend the approach against its competi-
According to Evans, one has a “fundamental Idea” of a kind of object – say, G’s – if one one has a “general
conception of the way in which G’s are distinguished from one another, and from all other things” (Evans, Varieties, p. 109). Thus one has a fundamental Idea of a particular G if one has “a specific answer to the question
‘What differentiates that object from others?’, of the kind appropriate to objects of that sort” (p. 107). Thus,
having a fundamental idea of a particular object “constitutes, by definition, distinguishing knowledge of the
object” (p. 107).
Brie Gertler, Self-Knowledge, London, New York: Routledge 2011, chap. 6.
tors, but to use it for developing an answer to our guiding question of what it is to know persons other than oneself.15
10. Self-knowledge is, according to a tradition as old as philosophy itself, a constitutive trait
of rational beings. Rationality comes in two forms, namely as practical and theoretical reason.
Being a reasoner therefore means being an agent and a believer, or, to put it differently, to
have the capacity to rationally determine one’s will and one’s understanding. An action is an
act of the will, resting on practical reasoning, whereas a judgment is an act of the understanding, resting on theoretical reasoning. Actions and judgments are both acts of thought (though,
of course, not necessarily thoughtful acts). Both are things someone does. But both are also
things someone does. Something is an action in the full sense of the term only if the agent is
conscious of it as being her action, if she attributes it, as the action it is, to herself.16 This in
turn implies that an agent must, in order to be an agent at all, be conscious of herself as an
agent. But the same goes for judging. A believer is only believing something in the full sense
of the term if she is, under some description, conscious of a judgment as being her judgment,
if she attributes it to herself.17 This in turn implies that a believer must, in order to be a believer at all, be conscious of herself as a believer. As Kant famously said: It must be possible for
The following sections are heavily indebted to Sebastian Rödl’s account in his Self-Consciousness, Cambridge
(Mass.), London: Harvard University Press 2007. Variations of the rationalist approach from which I have
learned are also spelled out by Akeel Bilgrami, Self-Knowledge and Resentment, Cambridge (Mass.), London:
Harvard University Press 2004; Richard Moran, Authority and Estrangement. An Essay on Self-Knowledge,
Princeton: Princeton University Press 2001; and in the writings of Tyler Burge collected in his Cognition
through Understanding: Self-Knowledge, Interlocution, Reasoning, Reflection, Philosophical Papers, vol. 3,
Oxford: Oxford University Press 2013.
Cf. Rödl, Self-Consciousness, chaps. 2 and 3. This is the fundamental case. There is room for fine distinctions
of more or less derivative cases here. Of course, I may do things unintentionally, i.e., without knowing it. Thus
in intentionally—i.e., self-knowingly, citing this as what one is doing—marrying Jocasta, one may unintentionally—without knowing it—marry one’s own mother (if one happens to be the nescient son of Jocasta). The latter is
a description of the same action, though not the one the acting subject would (presumably) give. It is still the
description of an action because what happens is still intentional—self-known by the agent—under some description.
The same point applies as in the footnote above: This is the fundamental case. In distinguishing various de-
rivative cases, one may be said to believe things which one does not know one believes, i.e. under specifications
of content which one would not endorse. Thus in self-knowingly acknowledging that Jocasta is attractive, one
might, in a certain, non-fundamental sense, incur a belief without one’s noticing it, namely the belief that one’s
mother is attractive, if Jocasta is in fact one’s mother, even though one would of course not be willing to selfconsciously acknowledge this thought.
the “I think” to accompany all my representations if they are to be my representations.18 Action and belief are inherently, essentially, self-conscious.
11. But what, we must ask, does it mean to say that the reasoner must be “conscious of herself” as a reasoner? It seems that she must know her own thoughts and deeds and that she
must know them as being hers. But this is not sufficient, as a long debate in the philosophy of
mind and language has established. For saying that the reasoner must know her thoughts as
being hers just says that she must know to whom they belong. And this would not suffice to
make her thoughts self-known to her in the way that is necessary for being self-conscious.
Rather, in knowing her thoughts as hers, she must know of herself in a unique way, namely,
as “I”.19
12. What does it mean to know oneself as “I”? One idea, famously criticized by Wittgenstein
in the Blue Book, is to understand “I” as a tool of reflexive deixis, identical in sense to an expression like “this speaker”, accompanied by a gesture of pointing to oneself. In such uses of
“I” (Wittgenstein calls them uses of “I” “as object”)20, I predicate something of myself in essentially the same way as others do, by referring to me “from the outside”, as it were, as one
object in the world amongst a manifold of objects. This means, however, that in judgments
including such a use of “I”, two components can be distinguished: (1) the judgment that this x
is F, (2) the separate judgment that I am this x. For that reason, there is a logical possibility of
misidentifying myself in judgments of such a form. Imagine a costume party in a ball room
whose walls are covered with mirrors, to which we all decided to go as Star Wars Imperial
Storm Troopers. If I see someone bleeding from under his costume in one of the mirrors, I
may take myself to be this person, and cry out: “I am bleeding!” But in such a case, it makes
sense to ask me whether I am sure that it is me who is bleeding. And indeed it may turn out
that I mistook someone else’s reflection in the mirror for mine, and that I am unharmed. (On
the other hand, I may see my own reflection in the mirror without realizing that it is mine.)
13. All this is different, however, in the fundamental use of “I” (or, as Wittgenstein says, use
of “I” “as subject”). These are the uses of “I” occurring in judgments in which I articulate
Cf. Immanuel Kant, Critique of Pure Reason, B 131 sq.
In what follows, I am drawing on insights discussed by, among others, G.E.M. Anscombe, “The First Person”,
in: Samuel Guttenplan (ed.), Mind and Language, Oxford: Clarendon Press 1975, pp. 45-65; Hector-Neri Castaneda, “He. A Study in the Logic of Self-Consciousness”, in: Ratio 7 (1966), pp. 130-57; Sidney Shoemaker,
“Self-Reference and Self-Awareness”, in: The Journal of Philosophy 65 (1966), pp. 555-567.
Cf. Ludwig Wittgenstein, The Blue and the Brown Books, Oxford: Blackwell 1969, pp. 66-70.
myself as an agent and a believer, judgments like “I believe that p” or “I will do F”. Such
judgments cannot be split up in two components, “Someone (this person) believes that p” or
“Someone (this person) will do F”, plus the independent judgment that I am this someone.
Therefore, such judgments are, as some have said, immune to error through misidentification
of their logical subject. This can be brought out by focussing on the fact that in such cases (as
opposed to the kind of case discussed above) it would be absurd (or indicative of some serious
intellectual disorder) for me to ask myself “Someone around here wants to do F. Is it me?”,
and equally absurd (or again, implying the allegation of a serious intellectual disorder) for
others to ask me “Are you sure that it’s you who believes that p?”
14. These well-known considerations show that knowing oneself as “I” constitutes a special
first-personal way for a reasoner to know herself, irreducible to any kind of internalized thirdpersonal knowledge. And this chimes well with an old and venerable idea: that the way in
which I know myself as “I” —self-knowledge—is different from any other way of knowing
me. It has traditionally been characterized by a special kind of authority or certainty. But we
are now in a position to understand what the real ground of this specialness is. It is not based
on a special, miraculously reliable faculty of inner vision (introspection) with which I observe
myself and detect mental states within me, as some philosophers have thought.21 In saying
what I believe or intend to do, I do not give reports about what I observe happening in me. I
am not answering the empirical question which beliefs and intentions I, as a matter of fact,
happen to detect in my mind. Rather, in saying what I believe or intend to do, I am looking
outwards, answering the normative or deliberative question what to believe and what to do.
Using “I”, I do not detect, I determine these things, and thus determine myself—as someone
who wills and believes such-and-such—in such acts of thought. Every rational determination
is a self-determination. In the fundamental sense, I know of my actions not by observing
them, but by acting, and I know of my judgments not by observing them, but by judging. This
is how I know myself referring to me as “I”: as a self-determining reasoner, an agent and a
believer. This knowledge is knowledge not from receptivity (from the capacity of empirical
observation), but from the capacity for practical and theoretical self-determination, which is
essentially self-conscious. Kant called this kind of knowledge “spontaneous” or “from spontaneity”. To conclude: The fundamental form of knowledge of oneself as a person—selfknowledge—is knowledge from spontaneity.
For a penchant critique of this idea, cf. Sydney Shoemaker, “Self-Knowledge and ‘Inner Sense’”, in: Philoso-
phy and Phenomenological Research 54 (1994), pp. 249-290.
III. The everyday conception of what it is to know someone “personally” determines this form
of knowledge as knowledge by interaction, and distinguishes it from knowledge by acquaintance
15. The general thesis I am going to develop in the rest of this paper is that what I said about
the fundamental form of knowledge of oneself as a person—namely, that it is knowledge from
spontaneity—is true of the fundamental form of knowledge of persons in general, hence also
of the fundamental form of knowledge of other persons. Self-knowledge and knowledge of
others are fundamentally and essentially of the same form. The fundamental way of
knowledge of a second person (the kind of knowledge we’re after) is knowledge from joint
spontaneity, that is, not by acting on and thinking about her, but by interacting and thinking
with her.
16. Before trying to vindicate this bold claim, I want to garner some initial plausibility for it
by showing that it in fact chimes well with some features of what I take to an uncontroversial
everyday notion of what it is to know someone “personally”, as they say. I take it that “I know
her personally” is a colloquial expression, and that most people, when asked what they mean
by it, would reply that they mean that they have actually encountered the someone in question. Now this may sound much like the explanation of the term “knowledge by acquaintance”
we rehearsed above, where we explained “being acquainted with O” as “having encountered
O in one’s perceptual experience”. But in the case of knowing someone personally, “having
encountered O” (O now standing for a person) clearly means something altogether different: it
includes the idea of interaction between the knower and the known, for example that they
have, on at least one occasion, spoken to each other. I believe that it is in line with the everyday usage of the expression to maintain that I cannot be said to know someone personally if I
have just, on some occasion, perceived the someone in question, but not interacted with her in
any way. Acquaintance, clearly, is not enough, no matter how large in scale it may be. For no
matter how long I have known someone by acquaintance and how many facts I come to know
about her based on acquaintance, this will not add up to knowing her personally.
17. This last contention might appear to be unwarranted. One may want to object that the alleged difference between knowing someone by acquaintance and knowing her personally (qua
knowledge by interaction), the objection goes, is really just a quantitiave difference. To ac11
count for it, it is sufficient to point out that in acting and speaking together with someone I see
what she does and hear what she says. Thus interacting with someone just gives me plenty of
opportunity to acquire knowledge by acquaintance of her. No wonder then that interacting
with someone will usually result in my knowledge of her being, all in all, broader, deeper, and
richer than it would have been if I had just observed her from behind my curtains. Thus, socalled knowledge by interaction would turn out to be ordinary empirical knowledge based on
acquaintance, just more of it. Such a reductive account of knowing someone personally would
be somewhat reminiscent of an account of the form of self-knowledge that is often ascribed to
Gilbert Ryle.22 Ryle is read as denying that there is a difference in form between the
knowledge I have of my own mental states and the knowledge of anyone else’s mental states.
It is just that I am, concerning myself, in a privileged observer’s position, since I cannot fail,
under normal circumstances, to note the contents of my own mind. The objection I am considering at present transposes this Rylean idea from the first/third-person case to the second/third-person case.
18. I believe, however, that the quasi-Rylean objection is mistaken. Here is a little thought
experiment to bring this out. Imagine that you have been secretly observed all your life, day
and night. Nothing escapes the secret observer’s attention and nothing of what has been observed is ever forgotten. The secret observer knows facts about you which you never knew
because you were too young to apprehend them, or which you have long since forgotten because they were so insignificant. Think of the secret observer as the Tower Society from Goethe’s Wilhelm Meister, equipped with GPS, CCTV, and unlimited disk storage. Or, better still,
think of the situation as alike to the one presented in Peter Weir’s motion picture The Truman
Show (1998). The protagonist, Truman, lives a quiet and happy life in a remote coastal town
which he has never left. He has no idea that the town and the coast are but a huge film set
equipped with thousands of hidden cameras, that the sky above him is not the sky but the roof
of a gigantic hall that has been erected around and above the set, and that his entire life is the
topic of a celebrated reality TV series that shows the real life of an individual—Truman’s
life—from his first day unto his last. The people around Truman, including his professed
friends and relatives, are actors following the instructions of a God-like director who goes by
the name of Christof, and, together with his crew, steers Truman’s entire world from an invisible control room.
Cf. Gilbert Ryle, The Concept of Mind, Chicago: University of Chicago Press 1984, pp. 194 sq.
19. The movie tells the story how Truman finds out, after all, that his world is fake. In the
end, he captures a rowing boat and sets out to the artificial sea, defying the artificial storm the
director unleashes in order to protect the grand illusion from being uncovered, until he suddenly bumps into the blue-painted outer wall of the film studio and discovers a door which
leads to the outside. In this situation, the director decides to reveal himself to Truman (the
religious overtones in this scene should be apparent) in an attempt to persuade him not to step
over the threshold that separates his artificial world from the real one. At the climax of their
ensuing exchange, the director speaks the following words to Truman: „Truman, I've watched
you your whole life. I saw you take your first step, your first word, your first kiss. I know you,
better than you know yourself. You're not going to walk out that door.“ Truman, however,
rejects Christof’s claim to know him out of hand. Despite the director’s infinite empirical
knowledge about him, Truman insists that Christof does not know him. And I think that most
people will share the feeling that there is something deeply right in Truman’s reaction. There
is a sense in which the director, no matter how vast his empirical knowledge of Truman is,
does not know him, indeed cannot even know him, as a matter of principle. Again, acquaintance is not enough. But what is missing cannot be a matter of the director needing more
knowledge by acquaintance of Truman, for ex hypothesi he knows everything there is to
know.23 What this little story shows, I believe, is that our everyday conception of what it is to
know someone “personally” distinguishes this form of knowledge from knowledge by acquaintance. Acquaintance (presence to someone’s perceptual experience) is not the fundamental form of knowledge of others.
IV. The fundamental form of knowledge of others is knowledge by interaction, which is
knowledge from joint spontaneity within a second-personal nexus of two persons performing
one act together
20. So what is the fundamental form of knowledge of others? As I already made clear, I think
that the examination of our everyday notion of knowing someone personally clearly points to
Truman himself, by the way, although he is entitled to reject the director’s claim to know him, has no convinc-
ing idea about why he is so entitled, as can be seen from his actual answer, which shows him to be a faithful
Cartesian: “You never had a camera in my head!” But this fails as a justification for his refusal, since the people
we do credit with knowing us don’t have cameras in our heads either.
an answer. In order to count as knowing someone personally, I minimally must have encountered the someone in question, not in my perceptual experience, but in interaction with her, on
at least one occasion. The fundamental form of knowledge of others is knowledge by interaction, on at least one occasion, minimally. If this is true—and I believe it is—the fundamental
form of knowledge of others will be explained by establishing an account of the form of
knowledge that is knowledge by interaction, for it is in interaction that I encounter the other as
a second person.
21. So what is an interaction? I take an encounter between A and B to be an interaction if its
form is constituted by its including an act of address and an act of response within the overarching unity of one single act. In such an encounter, the two subjects are bound to each other
by a practical nexus of one act they perform together. Michael Thompson calls the form of
such a nexus “bipolar”.24 Let me give an example.25 If somebody (A) sells something (X),
there must be a second person (B) who buys X. One person alone can neither buy nor sell anything – selling is necessarily to somebody other than oneself, buying is necessarily from
somebody other than oneself. But “selling” and “buying” do not designate two acts that A and
B perform independently. Rather, there is just one act which is done by A and B together. It is
not the case that there are two acts being performed: first A sells X (by abandoning it), and
then (as a second act) B buys it. This way of describing what is going on makes the idea of
buying and selling incomprehensible. For as long as B hasn’t bought X, A has not sold it. She
may have abandoned it, and she may even have abandoned it to B, but that does not constitute
an act of selling. So A hasn’t sold until B has stepped in to buy. But if, ex hypothesi, A has
already abandoned X before B buys it, X does not belong to A anymore, so B cannot buy it
from her. It is also not the case that A and B perform two independent acts exactly simultaneously in a sort of extremely precise choreography (presumably very difficult to coordinate,
demanding a lot of practice). Rather, that A sells X to B and that B buys X from A are two
ways of characterizing one and the same act, one and the same fact.26
Cf. Michael Thompson, “What Is It to Wrong Someone? A Puzzle about Justice”, in: R. Jay Wallace, Philip
Pettit, Samuel Scheffler, Michael Smith (eds.), Reason and Value: Themes from the Moral Philosophy of Joseph
Raz, Oxford: Oxford University Press 2004, pp. 334-384, here p. 335.
The example is Kant’s in §§ 18-19 of his Metaphysics of Morals, AA 6, pp. 271-273, translated in Immanuel
Kant, Practical Philosophy (Mary J. Gregor, ed.), The Cambridge Edition of the Works of Immanuel Kant,
Cambridge: Cambridge University Press 1999, pp 421-423.
This is not to deny that we can make distinctions between other kinds of acts within the act of buying/selling –
for example, A handing over X over the counter to B and B paying afterwards, or A signing a purchase contract
22. The act of selling something to someone (which is the act of buying something from
someone) requires that the someone to whom X is sold is the same one as he who buys X from
someone, who must in turn be the same one as he who sells X. Otherwise, what is going on is
not an act of selling and buying. Therefore, the very identity of this act includes the person to
whom X is sold.27 This is what it means to say that such an act is addressed: It is addressed to
someone in particular, and the idea of that particular someone is part of the identity of the act
itself. I cannot sell anything to anyone except to the one who buys it from me. If I plan to sell
X to B, but in handing it over inadvertently put it in C’s hands who has maliciously pushed B
to the side and quickly gives me some money in return, then X has not been sold to C and C
has not bought it. It was not, so to speak, for him to buy. (Compare: If I order B to do Y, but
instead of her, C does what I ordered B to do, then C has not obeyed my order, even if she did
what she did because she overheard my issuing the order to B. She cannot have obeyed my
order, since I did not order her. My order was not for her to obey.)
23. An interaction of the form I am considering is, as we saw, one act, not a conjunction of
two independent acts. It is an act which binds together two subjects in such a way that a sort
of “normative current”, as Thompson puts it, circulates between them, as between two poles.
Let us call them the positive and the negative pole, respectively. (This is a purely formal characterization). In the case of selling/buying, we have one act for two subjects, one being the
giving and the other being the receiving pole. But it is important to note that the “receiving”
part is not a passive object of the giver’s action. B does not receive the good she buys like
someone may receive a slap in the face. Slapping someone is an act for one subject. It does
involve a second subject, but only as object. Hence it does not constitute an interaction in my
sense. In an interaction, the person at the receiving pole is herself the co-doer of one, joint
act. The act must involve her not as object, but as self-conscious agent. She must be active in
it herself, to complete the circuit of normative current running through the act. This is what it
means to say that the unity of such an act must be a unity of address and response. Therefore,
an act of this form is necessarily performed by two persons together: It is an act of joint spon-
first, and B after him. But these individual acts are only intelligible as the acts they are because they are elements
of the joint act of buying/selling.
It might be objected that this is not true, that I can just intend to sell X, no matter to whom. But this undirected
action which is not addressed to someone in particular is not an act of selling, but at best an act of offering for
taneity. If one of them refuses to take part in it, the act in its entirety is not brought off.28 This
means that the bipolar nexus uniting A and B must be, in the fundamental case of genuine
interaction, be a nexus which unites A and B as self-conscious agents. This self-conscious
form of a bipolar practical nexus is what we call a second-personal nexus. Thus the fundamental form of knowing others, knowledge by interaction, is knowledge from spontaneity,
more precisely: from joint spontaneity within a second-personal nexus of two persons performing one act together.29
24. The idea that the fundamental form of knowing others is knowledge from spontaneity (not
from receptivity) may seem insane. Can I not know what someone does by (secretly) watching her, and can I not know what she thinks by hearing her talking, without any interaction
taking place? And is this not a case of receptive (as opposed to spontaneous) knowledge? Of
course. And I see no reason to deny that these are genuine cases of receptive knowledge, of
perceiving what someone thinks, feels or does.30 The point is just that such receptive
knowledge is not the fundamental form of knowledge of others, for it presupposes a grasp of
the other as other, that is, as a second person, a fellow agent. And this concept cannot come to
us from perception. It can only come to us from interaction itself, by receiving the other’s
“summons” (to use Fichte’s term) to respond to her by engaging in a second-personal act.
Once the concept of a person has been acquired by being initiated in such relationships, it can
then be employed, in a logically derivative way, outside such relationships, for example, in
empirical observation of others.
This does not mean that no act has been brought of, that nothing has been done. It is just that in such a case, an
entirely different has been performed, e.g., a purchase contract has been signed by only one of the parties. But
this is not, somehow, half an act of selling, but simply no act of selling at all. If A, the putative seller, signs the
purchase contract, but B, the putative buyer, suddenly dies before she can sign it too, it would be nonsense to say
“Well, poor B didn’t have the chance to buy X anymore, but how lucky A is that she at least managed to sell it!”
Note that there are a lot of actions which are intrinsically social, but not addressed to someone in particular
and hence not bipolar-interactional and second-personal in the sense discussed here. Different accounts are
called for of these different forms of social action, thought, and intentionality. They are the topic of the huge
literature on collective intentionality, into which I cannot go here.
Rödl, however, appears to be under the impression that he would have to explain away this kind of talk; cf.
Rödl, Self-Consciousness, pp. 178 sq.
V. The fundamental form of knowledge of another person is knowledge of her as as “you”
(second person knowledge), which is essentially knowledge of the same form as knowledge
of oneself as “I” (first person knowledge)
25. We said that it is part of the very identity of a second-personal act that it is addressed to
someone in particular. The idea of that particular second person is an element of the very idea
of the act itself. We also said that action as such, in the fundamental case, is essentially selfconscious. If someone performs an action, she knows from spontaneity what she is doing.
However, if knowing what one is doing includes, in the case of a second-personal act,
knowledge of the second person with whom one interacts, it follows that in performing a joint
second-personal act, I know the someone with whom I am performing it together, and know
her from spontaneity – otherwise I would quite literally not know what I am doing and therefore would not be doing it. But how is that possible? How can I know someone as the one
with whom I am interacting, just by interacting with her, i.e., spontaneously? What form of
conceptual representation makes this form of knowledge possible? Answering this question
will answer our question about the fundamental form of knowledge of others.31
26. In section II we saw that the conceptual representation that sustains spontaneous
knowledge of oneself as a reasoner, an agent and a believer, is the first person pronoun, “I”.
Our present question is which form of conceptual representation sustains having the same
kind of knowledge of someone else as the one with whom I am bound together in a practical
nexus of a joint second-personal act. The analogy with first person thought points to the answer to this question: The conceptual representation that makes such knowledge possible is
the second person pronoun, “you”. We said that second-personal acts are essentially addressed. I take it to require no argument that calling someone “you” just is addressing her and
that there is no meaningful “you”-saying outside acts of address, of interaction.32 Furthermore, saying “you” is the only way of addressing a thought about someone to that someone.
Being presented to N.N., I may say that “N.N. is F” or that “This person is F”, and N.N. may
The following considerations are, almost in their entirety, Sebastian Rödl’s, as developed in his Self-
Consciousness, chap. 6, and his “You & I”, unpublished manuscript.
Obviously, the interaction does not have to be face to face and hence does not have to include acquaintance, i.
e. encountering each other in one’s perceptual experience. Trivially, I can address someone as “you” in a letter.
More interesting non-fundamental cases beyond the scope of the present discussion include addressing as “you”
the readers of a text or the finder of a message in a bottle.
understand that I am talking about her by knowing that “N.N. = I” or “This person = I”. But
she may fail to grasp that I am talking to her as well. She may consume my speech act passively, without it constituting any kind of normative nexus between us. Only by saying “You
are F” does my thinking as such address her as the someone who figures in it.
27. So we can say that conceptually representing someone as “you” makes her known to me
as the one whom my act addresses. Just as in the case of “I”, this constitutes a way of knowing that is irreducibly different from any other. Just as “I” is not semantically equivalent to
“the person now speaking”, “you” is not semantically equivalent to “the person to whom I am
now speaking”.
In order to bring this out, let us return for a moment to the costume party in
the ball room whose walls are clad with mirrors. In such a context, we can immediately see
that “I will take you home” expresses a thought which is different from the thought expressed
by “I will take home the person to whom I am now speaking” (where the latter expression is
accompanied by a pointing gesture to the person in question). For the addressee can fully understand the latter expression without understanding the former. This is the case if she sees
the person whom the speaker points to in the mirror and thus knows that “the person to whom
the speaker is now speaking = that person” (being perceptually attuned to the right person),
but fails to know (due to those Star Wars Imperial Stormtrooper costumes) that “that person”
in the mirror is her.34 That is not something she would need to know in order to understand
what is said by “I will take home the person to whom I am now speaking”. However, she
would need to understand it in order to know what is said by “I will take you home”.
28. We established that an interaction is an act of addressing someone as a second person, that
is, as a fellow agent and believer. And we pointed out that this is just to say that addressing
her means addressing her as someone capable of referring to herself as “I”. Therefore, addressing someone is not an act of pointing at her, sustained by a relation of perceptual acquaintance, as in a case of demonstrative reference, but of summoning, that is calling upon
her to respond as a free, self-conscious being, by joining in interaction with the addresser,
which is, by saying “I”, by committing and determining herself. If our account is correct, this
structure of reciprocity should be sustained by the grammar of “you” and “I”. And indeed the
mirror case discussed above shows that the only way to fully understand a “you”-reference
The view that these two expressions would be semantically equivalent is articulated by Richard G. Heck, “Do
Demonstratives Have Senses?”, in: Philosopher’s Imprint 2 (2002), pp. 1-33, here p. 12.
This source of misunderstanding and others of a related kind are what make the fancy dress ball such an attrac-
tive setting for romantic comedies.
addressed at oneself is to understand it by knowing that “you = I”. This means to understand it
in a way which is uniquely possible for the addressee of the act being the act it is, namely an
act that is as such directed at her, that is hers to participate in. Thus, understanding someone
else’s “you”-thought addressed to oneself is understanding that one is being called upon to
respond to the thought by way of a free, self-conscious commitment, to close the circuit of
normative current in the second-personal nexus of addresser and addressee. And understanding this already means that the circuit is closed, for even an act of refusal then is a way of letting the normative current flow. (Refusal is a self-conscious act and hence of an altogether
different form than simply remaining deaf to the summons.)
29. According to the view presented here, knowledge of someone as “you” is knowledge from
(joint) spontaneity, not from receptivity. My fundamental way of knowing you is not
knowledge from perception. To this one might want to object that every so often I look for
someone and, when I have spotted her, address her by pointing at her, saying: “You!” Is it not
obvious that saying “you” is very often accompanied by a gesture of pointing? And isn’t that
a case of demonstrative reference, sustained by a perceptual relation to an object, hence receptive knowledge? But the observation, though correct, does not show that knowledge of who it
is that I am addressing is, in the fundamental case, knowledge from perceiving the someone in
question. It does show that having receptive knowledge (knowledge based on perception) of
people is an enabling condition for joint action, as it is for action in general. Presumably, I
could not play the violin if I could not hear. That does not show that I know from hearing myself play that it is me who is playing. Rather, I know this spontaneously, simply by playing.
But so it is with addressing someone. I do not have to see you in order to know that it is you
whom I address. Rather, I intend to address you, and my knowledge that it is you to whom I
want to turn myself is fixed by my intending this. This can be made vivid by considering the
difference with demonstrative reference: In the case of demonstrative reference, it is precisely
by mediation of perception that I know the object to which I refer. Therefore, the object which
I perceive fixes what it is that my thought is about. If I think of an object presented to me in
perception as “this object”, every thought I think about it will be about this very object, whatever it is. (If it turns out that there is no object, because I was only under the illusion to be
presented with one, there is no thought of mine.) Thus, if I think of someone perceptually present to me as “this man”, and say of him, for example, that he is Peter Weir, I will have predicated this of this man presented to me. If it turns out that he is not Peter Weir, I have to admit
my mistake by saying: “I said indeed that you are Peter Weir. But I see now that I was mis19
taken in saying this.” By contrast, if I think of someone I want to address, for example, by
saying “You are driving me insane!”, the reference of “you” will be fixed by my intention
whom to address. If I then happen to mistake you for someone else who is visually presented
to me and looks similar to you, resulting in my speaking these words to the wrong person, that
does not make it the case that I have thereby spoken about the person who is standing before
me. I would certainly apologize to this person. But I would do so precisely not by saying “I
said indeed that you are driving me insane. But I can see now that I was mistaken,” but rather
by saying, “I was not talking about you!”
30. We saw that I can only know what is said by expressing a “you”-thought by framing an
“I”-thought, saying “I” where you said “you”. We can now complete this line of thought by
shawing that the same is true vice versa: In order to fully understand your “you”-thought, I
have to say “you” where you said “I”. For, we said that I do not know what is being said by
the expression of a “you”-thought if I do not know it as something that is addressed to me,
that is for me to respond to. But a response is an act of address, too. A response to a call is of
necessity to the one who initiated the call, otherwise it fails to be a response. Hence we can
repeat the considerations rehearsed above in order to show that everything we said about the
act of address depending on framing a “you”-thought, the understanding of which depends on
the framing of an “I”-thought on part of the addressee, equally applies to the backflow of
normative current in the second-personal nexus of joint action from addressee to addresser. I
know you as “you” only if I know you as someone who knows herself as “I”, and you know
yourself as “I” only if you know yourself as someone whom I know as “you”, and vice versa.35
31. It follows that, even though the concrete second-personal act in question may be asymmetrical in that A and B occupy two different poles in it (e.g., have different positions vis à vis
their joint act), the logical form of their practical nexus rests on a deeper symmetry between
the poles. A second-personal nexus is, by its very form, symmetrical. But this implies that the
terms of a second-personal nexus must be identical in logical form, i.e. they must be persons
with the capacity to grasp themselves as an “I” that is a “you” for others, and the other as a
“you” that is an “I” to itself. Thus, in a second-personal nexus, the other is represented in the
same form as I am. Hence, according to the account presented here, knowledge of others is
knowledge of the same form as knowledge of oneself. The fundamental form of knowledge of
a person, which is knowing her as “you” (second-personal knowledge), is essentially the same
Cf. Rödl, Self-Consciousness, p. 190.
as the fundamental form of self-knowledge, which is knowing oneself as “I” (first-personal
knowledge). “You” and “I”, sustaining second person knowledge and first person knowledge
respectively, are but two guises of a single generic form of knowledge of persons as such.
VI. In knowing each other as “you” and “I”, two persons united in a second-personal nexus
share in one act of grasping the same thought from two perspectives: they share the same second-personal knowledge
32. We are now in a position to explain how A and B, in interacting, representing each other
as “I” and “you”, thereby share one and the same knowledge from spontaneity. If A and B are
united in a second-personal nexus, the sentence (I)R(you), as uttered by A, and the sentence
(you)R(me), as uttered by B, express the same non-accidentally true thought, a thought that is
shared between A and B in their spontaneous knowledge by interaction. In order to see this,
we must not let ourselves be confused by the two sentences being composed of different linguistic expressions, “I” and “you”. Frege maintained that it is possible that two sentences with
different component expressions express the same thought. According to him, the sentences
“At Plataea the Greeks defeated the Persians” and “At Plataea the Persians were defeated by
the Greeks” have the same conceptual content, such that, thinking either one of them, we
grasp the same thought.36 Why is this but one thought? Because, according to Frege’s criterion of difference for thoughts, there is a difference in content between two sentences if and
only if it is possible for a subject to rationally assent to the one and dissent from the other at
the same time. But it is hard to see how this should be possible in the present case. “Defeating/being defeated”, just like “buying/selling”, is one and the same bipolar relation between A
and B, not two such relations. Therefore, anyone who knows or understands the relation that
at Plataea, the Greeks defeated the Persians, cannot fail to know or understand that at Plataea,
the Persians were defeated by the Greeks. For it is not clear what “knowing that A defeated B”
should be if it did not include “knowing that B was defeated by A”, and vice versa.37 The dif-
Gottlob Frege, “Begriffsschrift. Eine der arithmetischen nachgebildete Formelsprache des reinen Denkens”, in
his Begriffsschrift und andere Aufsätze, Hildesheim: Olms 2007, § 3.
This is different in the familiar kind of evening star / morning star case. In knowing or understanding “The
morning star is a planet of the solar system”, a thought is grasped which is different from the thought grasped by
ference between the two sentences is just a switch in means of linguistic expression, a switch
from the active to the passive voice, but not a difference in the thought grasped. It is always
possible to read a relational expression of the form (a)R(b) from left to right or from right to
left, so to speak. It expresses the same thought even if a language has different expressions for
it, depending on the direction of the reading (as in the case of buying/selling, lending/borrowing, and so on).
33. Frege also claims that sometimes, one has to use different linguistic expressions in order
to grasp the same thought twice.38 In order to grasp the same demonstrative thought several
times from different positions in time, one has to use different means of expression. If I
grasped a thought yesterday by thinking “Today the weather is fine”, I can only grasp the
same thought again today by thinking “Yesterday the weather was fine”. As Evans writes:
“Frege’s idea is that being in the same epistemic state may require different things of us at
different times; the changing circumstances force us to change in order to keep hold of a constant reference and a constant thought – we must run to keep still.”39 More importantly for
Evans, the same applies to grasping the same thought from different positions in space (which
will, since one can only occupy one position in space at a time, also be different positions in
time), as when we refer to a place we’re in by “It’s hot here”, and later, after we’ve moved, by
“It’s hot there”. Again, why do these acts express but one thought (rather than a sequence of
sufficiently related different thoughts)? Because, once more, no-one could count as understanding the concept “here” if she did not understand the concept “there” and its essential relatedness to “here”, nobody could count as understanding the concept “today” if she did not
understand “yesterday” and “tomorrow” and their essential relatedness, and so on. In the fundamental case, to grasp a demonstrative thought überhaupt is manifested by a capacity to
keep track of the thought while one is moving through time and space, by expressing the
thought differently depending on the space-time-position one occupies.40
34. Now let us return to the first example. As Frege said, anyone who knows that at Plataea,
the Greeks defeated the Persians, thereby knows that at Plataea, the Persians were defeated by
knowing or understanding “The evening star is a planet of the solar system”, since one can understand the expression “the morning star” without understanding the expression “the evening star” and vice versa.
Cf. Gottlob Frege, “Thoughts” (“Der Gedanke”), in his Collected Papers on Mathematics, Logic, and Philos-
ophy, Oxford: Blackwell 1984, pp. 351-372.
Evans, Varieties, p. 194.
That does not mean, of course, that one cannot lose track of the thought and fail to retain it in some cases.
the Greeks, for this is but one thought. The difference in expression is merely superficial. We
could, for every bipolar nexus, formalize this as saying that anyone who knows that
A[Nexus(+/-)]B thereby knows that B[Nexus(-/+)]A, because this is one and the same
thought, where “Nexus(+/-) and “Nexus(-/+)” stand for the two possible grammatical expressions (active/passive) of the same nexus (e.g., “defeat”/”being defeated”). One either understands both of these or does not understand either one. But now, all we have to do is to ask
how the very same thought which an observer grasps by thinking “At Plataea the Greeks defeated the Persians” or its passive counterpart must be expressed from within this nexus,
which is not only bipolar, but second-personal, insofar as the terms of the nexus are selfconscious subjects. Enter A, the ferocious Greek, and B, the crestfallen Persian. Now, when A
says to B, “We defeated you at Plataea “ or “You were defeated by us at Plataea”, B must take
this up, in talking to A, as “You defeated us at Plataea” or “We were defeated by you at Plataea”.41 But both pairs of expressions can be rendered as We[Nexus(+/-)]You and
You[Nexus(-/+)]We. Therefore, they express the same thought.42
35. Such thoughts are thoughts which, in order to be grasped at all, must be grasped from
more than one position, just like demonstrative thoughts involving temporal or local indexicals. Since the form of such a thought is bipolar, it must be grasped from exactly two positions. But since its form is not just bipolar, but second-personal, exhibiting a nexus whose
terms are self-conscious, these two positions are two self-conscious perspectives on the
thought which require different means of expression. The thought therefore must be grasped
from two perspectives in order to be grasped at all. But this is equivalent to saying that such
a thought is of a form which is either thought by two thinkers together, or not thought at all. I
manage to grasp an I/You-thought only if it is, at the same time, the You/I-thought which you
grasp. Michael Thompson has proposed the expression “propositional nexus” to designate
such a propositional attitude or act of thinking for two.43 If there is no-one to think such a
thought together with me (because I erroneously addressed the scarecrow in the garden), I
simply did not grasp such a thought, but only made a failed attempt at grasping one. There-
I am sure that the fact that this is a plural first person / second person thought makes matters somehow more
complicated, but not, I hope, for my present concerns.
This is a version of an argument Rödl employs in “You and I”.
Cf. Michael Thompson, “How Shall We Represent Propositional Attitudes and Propositional Nexuses? Some
Ill-Advised Experiments in Ideography”, unpublished manuscript.
fore, thinking about you as “you” (thought de te) in the fundamental case is a case of thinking
one and the same thought together with you (thought tecum).44
36. It follows that the fundamental form of knowing someone is as such a form of mutually
shared knowledge. Or, to put it simple: Knowing someone, fundamentally, is knowing each
other. This sets knowing someone by interaction apart from observational knowledge (by acquaintance) which is one-sided: If A knows B observationally inside and out, that does not as
such imply that B knows A observationally, at all. (Just think back to Truman and Christof.)
And we can show that the intrinsic sharedness of second-personal knowledge is of a kind that
is irreducible to observational knowledge even in cases where the latter is, in a sense, reciprocal. We saw that she sharedness of second-personal knowledge means, in the fundamental
case, that A and B know each other in exactly the same way—as I/you, you/I—as they share
the same knowledge from two perspectives. Precisely this is what is impossible in the case of
mutual observational knowledge, no matter how sophisticated. Here’s how to see this: Imagine a case in which A knows from observation everything there is to know about B, but, equally, B knows from observation everything there is to know about A. We might imagine a case
that is like Truman’s, except that here, Truman has somehow turned the tables on the director
and is now observing him, too. Now there is a range of facts concerning Truman which he
knows from spontaneity, referring to himself as “I” (these would be facts concerning what he
thinks and does), but which the director would know from observation, referring to Truman
(not as “you”, but) as “Truman”. Analogously, there is another range of facts concerning the
director which he would know from spontaneity (this being facts concerning what he thinks
and does), but which Truman would know from observation. But this means that their
knowledge of the relevant facts, and their respective judgments, would differ in form. Therefore, it would not be the same knowledge. This will not change if, unavoidably, they will at
some point start to observe one another observing one another. Thus Truman would know
from spontaneity that he observes the director whom he knows from observation to know that
Truman observes him. But the same fact would be known to the director in knowledge of a
different form: The director would know from observation that he is observed by Truman
whom he knows from observation to know that the director observes him. (The same goes for
the corresponding fact that the director observes Truman.) We might now add further clauses
specifying that Truman knows that the director knows that Truman knows, and that the director knows that Truman knows that the director knows, and so on ad infinitum, but all this pil44
I owe this formulation to a conversation with Günter Zöller.
ing up of higher level observational knowledge will not bridge the fundamental rupture in
form of knowledge at the ground level. What we have in such a case will always be two separate bodies of knowledge. We might say that these bodies of knowledge are, in a sense, reciprocal: they complement each other. But it is one thing to have two opposing bodies of reciprocal knowledge that complement each other, or to have one mutually shared body of
knowledge that is owned by two. Truman and the director get that only in the moment when
they start interacting by personally addressing each other as “you” and “I”. At that moment,
one could say, despite the huge scope of their reciprocal knowledge from observation, they
start from scratch as far as their personally knowing each other is concerned. In a certain way,
they meet like total strangers. What they have to do is to start talking to each other, i.e. interacting, and thereby get to know each other for the first time.
37. Personal knowledge, the kind of knowledge necessarily shared (in its first-personal and its
second-personal guises, respectively) between agents joined in a second-personal nexus of
joint action, involves the other not as the mute object of my knowledge, but as co-agent and
interlocutor, deliberating, examining, sharing the same thought with me. Thus, at the end of
our analysis, it turns out that we have to reject the seemingly innocent idea from which we
started, namely that in ascribing knowledge of a person to someone, we describe her as relating, in the manner of knowledge, to an object. This is wrong, and it is not put right by adding
that the object in question is of course an object of a distinctive kind, i. e., a person. Rather, it
turns out that what it is to know someone, in the fundamental sense, is not a matter of a subject relating to its object at all. Rather, if A and B are united in a second-personal nexus of a
joint act, they mutually relate to and know each other as the joint subject of this act. This is
the fundamental form of what it is to know someone.
VII. Concluding Remarks
38. I have completed my sketch of the proposed account of what it (minimally) is to know
someone. Let me conclude with a few tentative remarks concerning the question what philosophical consequences this account—if it could be reinforced and defended—would have. It
is telling that the topic of knowing others occurs in philosophy mainly in the guise of the socalled “problem of other minds”, which is commonly divided into a “conceptual” and an
“epistemological” problem. The “epistemological problem” concerns the question of how I
can know with any certainty what someone else is thinking or experiencing. The “conceptual
problem” however concerns the question of how I can even have concepts of mental states
that apply both to me and to the other, since—according to the received terms of debate in this
context—I predicate those concepts of myself on completely different grounds than I predicate them of others. Now the first thing to note is that the account of what it is to know someone sketched in this paper leaves no room to even set up the supposed conceptual problem of
other minds, as Sebastian Rödl points out.45 For it shows that, contrary to the empiricist preconceptions most participants to the debate simply take for granted, the form of my
knowledge of my own mental states and the form of my knowledge of your mental states are
fundamentally the same. There is no gap to be bridged.
39. The epistemological problem of other minds is usually framed in empiricist terms as well.
That is, the only form of knowledge ever considered in this debate is empirical knowledge,
knowledge based on acquaintance or observation. What is in question is just the best method
of accessing the object.46 We saw, however, that it is impossible to understand fundamental
knowledge of others in terms of an empiricist conception of knowledge. Since, on the basis of
the currently dominant empiricist preconceptions, no other conception of knowledge is available, critics who see the hopelessness of the empiricist conception of what it is to know someone tend to reject the concept of knowledge of others altogether, claiming that our relation “to
others in general is not one of knowing” 47, as Stanley Cavell puts it, but rather one of, say,
acknowledging, recognizing, or responding to the other. This reaction however threatens to
surrender the concept of knowledge to the empiricist. Our conception of knowledge of others
as second-personal knowledge from joint spontaneity can claim the advantage of reconciling
what is surely right in the reaction with our everyday talk about knowing and being known by
others. It does so precisely by outlining a form of knowledge acquiring which is not opposed
to, but rather is in itself an act of acknowledging, recognizing and responding to the other.
Cf. Rödl, Self-Consciousness, pp. 165 sq.
Monika Dullstein, in her contribution to this volume, shows how different approaches in the so-called “Theory
of Mind”-Debate invariably construe knowledge of other minds as constituted by one’s standing in a relation to
an object, be it an object of direct perception, a theoretical object, or something ready-to-hand with which to
have practical dealings. Cf. also her “The Second Person in the Theory of Mind Debate”, in: Review of Philosophy and Psychology 3 (2012), DOI: 10.1007/s13164-012-0095-2.
Stanley Cavell, The Claim of Reason. Wittgenstein, Skepticism, Morality, and Tragedy. Oxford: Oxford Uni-
versity Press 1979, p. 45. Cavell adds, “where knowing construes itself as being certain”, which makes his claim
40. This way of looking at things opens the prospect of an as yet unexplored branch of social
epistemology, centering on the question what it is and what it takes to know each other well.
As I said in the very beginning of this paper, knowledge of others is gradable, and I have offered but an account of the most basic instances of it. The argument in the previous sections
was meant to establish a minimal conception of what it is to know someone. It will have occurred to the reader that knowledge of others that fits this minimal conception is instantiated
by even the most minuscule and trivial everyday exchanges and interactions, like asking for
the time a total stranger on the street, or buying a snack from a street vendor one will most
likely never see again in one’s life. These are encounters that nobody in real life would cite as
cases of knowledge of others. In focusing on such cases, we seem to be lightyears away from
the kind of knowledge claims which I suggested play a vital role in our social lives. Still, I
believe that the form of these insignificant encounters is the seed of even the most formidable
forms of knowing each other, and that we can construct a full account of the more demanding
and more interesting cases by building on them.
41. As I said, this is not the place to even begin this task, but certainly one key factor in the
passage from cursory to deep cases of knowing someone is time. It is by continuous, multifaceted and intensive interaction that we get to know each other well, that we achieve—in the
best cases—that sort of intimacy and intuitive understanding which characterizes relations of
persons who have known each other for a very long time.48 On the other hand, the passage of
time is of course not sufficient for getting to know each other well. Something else that seems
to be required are virtues of knowing each other – honesty, truthfulness, attentiveness, but
also tact and discretion. There also appears to be something like an attitude, a willingness to
reveal oneself to a particular other, “sich zu erkennen zu geben”, as the Germans say—
literally: “to give oneself over to being known”—and a corresponding respect (or lack thereof) for the decision to reserve the right not to be known by a particular other. For, up to a certain point, I have a say in the matter of who shall know me second-personally. Consequently,
there is something to emphatic knowledge ascriptions like “You really know me inside and
out” that makes them sound more like granting a right to someone. By contrast, there is an
element of indignation in my rejecting someone else’s claim to know me as unrightful, as a
kind of epistemic injustice, by saying: “How dare you? You do not know me at all!” All this
This is where the figure of the friend and especially of the old friend, initially mentioned in the quote from
Jeremy Fantl’s article, as the paradigm instance of someone who knows me well, reappears on the horizon of our
would seem to entail that searching for knowledge about someone is an altogether different
enterprise than searching for knowledge about something. This is simply due to the fact that it
has to be done together with her. One gets to know each other better and better, not by observing one another and guessing each other’s mental states, but by living one’s lives together.49
I am grateful to Fabian Börchers, Amber Carpenter, Monika Dullstein, David Löwenstein, Jan Müller, Chris-
tian Skirke, and Anna Wehofsits for commenting upon even more preliminary versions of the present material. I
have also benefited from presenting and discussing parts of the paper at the Christian-Albrechts-Universität zu
Kiel and at the Humboldt Universität zu Berlin. My greatest debt is to the participants and convenors of the
Summer Institute (jointly sponsored by the National Humanities Center and the Wissenschaftskolleg zu Berlin)
on “The Second Person”. Rarely have I learned more or thought harder than by interacting with them.