Tiempo Klein
Tiempo Klein
Tiempo Klein
edited by
Wolfgang Klein
Ping Li
Mouton de Gruyter
Berlin · New York
Mouton de Gruyter (formerly Mouton, The Hague)
is a Division of Walter de Gruyter GmbH & Co. KG, Berlin.
앪
앝 Printed on acid-free paper which falls within the guidelines
of the ANSI to ensure permanence and durability.
ISBN 978-3-11-019581-1 hb
ISBN 978-3-11-019582-8 pb
쑔 Copyright 2009 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin.
All rights reserved, including those of translation into foreign languages. No part of this
book may be reproduced in any form or by any means, electronic or mechanical, including
photocopy, recording, or any information storage and retrieval system, without permission
in writing from the publisher.
Cover design: Frank Benno Junghanns, Berlin.
Printed in Germany.
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Wolfgang Klein and Ping Li
Concepts of time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Wolfgang Klein
Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Introduction
Wolfgang Klein and Ping Li
The ability to express time belongs to the most fundamental traits of human
communication. All human languages that we know of provide their speak-
ers with a range of lexical and grammatical devices to say when something
happened and how long it lasted, to say whether it happened, or will happen,
for the first time, regularly or very often, and to say whether some event or
state precedes, overlaps with or follows another event or state. These de-
vices include grammatical categories such as tense and aspect, certain fea-
tures in the lexical meaning of verbs, various types of temporal adverbials
and particles, but also discourse principles such as the maxim to tell events
in the order in which they happened.
In many languages, one of these devices, tense, is so deeply rooted in
the grammatical system that it is hardly possible to utter a sentence without
referring to time. This may be the reason why the study of how time is ex-
pressed in human languages has a strong bias towards tense and, to a
somewhat lesser extent, aspect – two categories which are often combined
to what traditional grammars usually call a “tense system”. This bias is un-
fortunate for two reasons. First, many languages have no inflectional mor-
phology at all, hence, no categories as tense and aspect in their grammatical
system (e.g., Chinese has no grammatical tense). This does not mean, of
course, that the speakers of those languages cannot indicate that something
is in the past, the present, or the future, or that something is on-going or
completed; they just use other means, for example particles or adverbials.
Second, in languages like Greek, English or French, in which tense and as-
pect are parts of the grammatical system, the expression of time is not con-
fined to these two devices. Typically, the temporal information which the
speaker wants to convey is encoded by a combination of various means,
including adverbials, inherent temporal features of the verb and discourse
principles. Hence, any real understanding of how the expression of time
works requires a somewhat broader perspective. This book tries to provide
such a perspective.
2 Wolfgang Klein and Ping Li
strates that their system is based on a very systematic and clever use of
various types of temporal anaphora. His analysis not only shows that tem-
porality can function very differently from what we are used to think; it also
challenges us to have a fresh look at those languages which are supposed to
be relatively well-described.
Whereas all natural languages have temporal expressions, formal lan-
guages – for example languages of logic or programming languages – typi-
cally lack such devices. But about forty years ago, philosophers and lin-
guists began to develop more complex formal systems which permit a
precise analysis of how, in natural languages, the meaning of compound
expressions results from the meaning of its components. Arnim von
Stechow’s chapter “Tenses in compositional semantics” illustrates this for
English; beyond tenses proper, he also examines lexical and grammatical
aspect.
We do not know in which way “human time” – that is, the way in which
we experience time and think about time – is shaped by our genetic endow-
ment, on the one hand, and by social and cultural experience, on the other.
But clearly, the way in which temporality is expressed must be learned,
since languages differ considerably in this regard, no matter how similar
the underlying temporal notions may be. In his chapter “Temporal expres-
sions in first and second language acquisition”, Yasuhiro Shirai gives a sur-
vey of what we know about these developmental processes, in particular
how the acquisition of the mother tongue differs from the acquisition of a
second or third language.
One reason why the study of temporality has made much less progress
than it could have done is a certain methodological narrowness, coupled
with an unbalanced diet of examples: by far most claims in the literature are
based on speaker’s intuitions of what certain forms – say he was sleeping
vs. he slept – mean, or on the study of occurrences of such forms in a text
corpus; in either case, the data have a strong bias towards utterances which
describe singular events in (real or fictituous) past. Both procedures, even if
extended to other text types and usages, have serious shortcomings. The
chapter “New perspectives in analysing aspectual distinctions across lan-
guages” by Christiane von Stutterheim, Mary Carroll and Wolfgang Klein
shows that other – and actually very simple – experimental methods can
give a much more reliable and differentiated picture of how speakers of
different languages use certain temporal forms. This is illustrated here for
aspect; but the methods can easily be extended to other temporal devices.
There is a clear difference between the temporal properties of a situation
(state, process, event) itself and the mental representation which the speaker
4 Wolfgang Klein and Ping Li
has of this situation, when he or she sets out to speak about it. It is the
speaker’s mental representation, rather than the situation itself, which is
crucial for the linguistic expression of time. The more or less subjective
“view” on the situation is closely connected to grammatical and lexical as-
pect. In the chapter “Verb aspect and the mental respresentation of situa-
tions”, Carol Madden and Todd Ferretti first examine the traditional classi-
fication of these two notions, and then, they present evidence from recent
psycholinguistic experiments on how speakers construct the discourse mo-
del that underlies their language production.
The concluding chapter “Computational modeling of the expression of
time” by Ping Li and Xiaowei Zhao is devoted to a very different way to
investigate temporality – computational approaches. The strong association
and interaction between lexical aspect and grammatical aspect, particularly
in the domain of first language acquisition, has previously led some re-
searchers to argue for innate semantic categories or prelinguistic predispo-
sitions. Ping Li and Xiaowei Zhao provide counter-evidence to this argu-
ment with simulations of aspect acquisition in a connectionist network,
DevLex-II. The simulation results indicate that the strong association be-
tween lexical aspect and grammatical aspect can emerge from dynamic
self-organization and Hebbian learning in dynamic computational systems,
therefore invalidating a priori assumptions about specific structures of in-
nate linguistic or conceptual knowledge. Thus, computational modeling is
not an aim in itself – it is a tool to verify or falsify existing claims on how
time is expressed, and to provide novel simulation data that can further in-
spire new empirical studies.
The expression of time marks an important human linguistic capacity,
and the study of it requires the joint efforts of linguists, psychologists, and
other cognitive scientists. It is our hope that this book serves as a catalyst
for future research that will elucidate the cognitive and linguistic processes
underlying the expression of time.
Concepts of time
Wolfgang Klein
Time has always been difficult to under-
stand, but in the twentieth century, our un-
understanding has become clearer.
J. R. Lucas (1999: 1)
1. Introduction
The experience of time and the need to adapt our life to it are as old as
mankind. The sun rises in the morning and sets in the evening, the moon
changes its position at regular intervals, plants and animals and humans
come into existence, grow, fade and pass away. We act here and now, but we
also remember having acted, and we plan and hope to act in the days ahead
of us. Some of these events, such as the coming and going of the seasons,
are cyclic, that is, they are repeated at intervals which we consider to be
equal. Other events are not assumed to be cyclic, such as our first love, the
birth of Jesus, or Grandmother’s death. All human cultures and societies of
whom we know have reacted in three ways to this temporal nature of expe-
rience:
– First, actions are planned and done accordingly – there is a time to plant
and a time to reap; a time to tear down, and a time to build; a time to
mourn, and a time to dance, as the Preacher has it in the Bible.
– Second, methods to measure time were invented. This is always done by
linking some event – the event whose duration we want to measure – to
some other type of events which are supposed to occur at regular inter-
vals, such the sequence of the seasons, the fall and rise of the sun, the
swing of a pendulum, the oscillation of a quartz crystal; the result are
calendars and clocks (Bruton 1993; Landes 1983; Richards 1998).
– Third, we speak about time. All human languages have developed nu-
merous devices to this end, and in some languages, the marking of time
is even close to mandatory. In English, as in all Indoeuropean languages,
the finite verb regularly expresses “tense” – that is, the sentence not only
describes some event, process, or state. It also places this situation into
the past, present, or future: we cannot say John be ill, thus leaving neu-
tral the time of the state thus described. We must say John was ill, John
6 Wolfgang Klein
is ill, John will be ill. Other languages, such as Chinese, have no manda-
tory marking of time. But this, of course, does not mean that they cannot
express time; they just use other means, such as adverbials like yester-
day, right now, or very soon, and they give their speakers full freedom
to indicate what happened when.
So, we all adapt our life to time; we use devices by which time is counted
and measured; and, above all, we speak about time. We know what it means
when someone says He will arrive tomorrow at five., The meeting has now
lasted for almost eleven hours., and Last february, I intended for the first
time to spend more than three hours per week in Pontefract. So, we do un-
derstand what time is. But what is it, then?
At this point it is common to quote St. Augustine, who, in the 11th book of
his Confessions, says:
His own way to overcome this clash between practical and theoretical un-
derstanding of time is that time is not in the things themselves but in our
soul. God, he says, is beyond time, and we get to know all things created by
him because he has endowed our soul with memory, experience, and ex-
pectation (see Flasch 2004). In other words, our soul – or our mind, as we
would probably say now –, is such that we experience the world as past,
present, and future.
St. Augustine’s theory of time is one of many within a rich stream of
thought that began with the first Greek philosophers and has steadily un-
folded over the millenia and over many disciplines – philosopy, physics,
biology, psychology, anthropology, linguistics, to mention but these. They
all deal partly with the same and partly with different aspects of time, the
result being a hardly permeable jungle of views, opinions, and theories. In
fact, the Augustinean question “Quid ergo est tempus?” has found so many
answers that one might as well say that there is no answer at all. Thus, the
idea that we could ever grasp “the essence of time” is perhaps futile; is is
doubtful that there is much more than a kind of family resemblance be-
tween a biologist’s, a phycisist’s, and a psychologist’s concept of time.
The aim of this chapter is not to unveil the “very nature” of time; it is
rather to prepare the ground for a basic understanding of how temporality is
Concepts of time 7
This section is a gaze into a jungle – into the rank growth of notions, ideas,
problems which have grown from a few germs laid in the Antiquity. At first
glance, it would appear to be hopeless to detect any structure in this jungle;
but in fact, there are a few recurrent themes which we will address in the
following section. It should be clear that this panorama is anything but ex-
1
The number of books on time is legion. The best general survey is to my mind
Whitrow 1980; it is, however, confined to time in philosophy, physics and biol-
ogy. Fraser 1987 is an easy and broad introduction by one of the best experts on
the study of time.
8 Wolfgang Klein
A. In Immanuel Kant’s Critique of pure reason (1781), time and space are
properties of human cognition – in fact, the two most fundamental cate-
gories of human cognition. They define the way in which our mind ex-
periences, and thinks about, the world. Time, in particular, is “die innere
Form der Anschauung”, (the inner form of intuition). It defines the way in
which we “intuit” external events and facts, such as the running of a horse
or the rotation of the earth, but also internal events, such the feeling of hun-
ger or grief. We cannot know whether time is “real”, that is, a property of
the world itself; our cognitive apparatus is such that the outer as well as the
inner world inevitably appear to us as structured by time.
B. In his influential article The unreality of time (1908), the British philo-
sopher John McTaggart argued that there are two types of event series, each
of which represents time: The “A-series” relates to the “earlier-later”-order,
to the mere succession of events, states, processes. In this sense, Aristotle
lived before St. Augustine, and Kant lived after St. Augustine. The “B-
series” relates to the difference between “past – present – future”. In contrast
to the A-series, it requires a particular vantage point, from which the events
are seen; this is the present moment – which, in turn, permanently shifts.
Under neither understanding is time “real”, argues McTaggart (see, e.g.,
Turetzky 1998).
2
In what is probably the oldest fragment of Greek philosophy we have, Anaxi-
mander says that the things, as they come into existence and perish, “pay their
debts to each other according to the order of time” (“kata tou chronou taxin”) – a
sentence of which no element is easily understandable (see Turetzky 1998: 6–8).
Concepts of time 9
Newton’s notion of real time is cryptic, perhaps because it has a strong re-
ligious background. As he states in the Scholium Generale of the second
edition of the Principia (1713) – an addition which particular famous for
Newton’ statement “Hypotheses non fingo (I don’t make up hypotheses)” –
he argues that time is an emanation of God, and God is in time (a position
which is in sharp contrast to St. Augustine’s, according to whom God is out
of time). It may well be that the tremendous success of Newtonian mechan-
ics is completely independent of his conception of “real” time. What is cru-
cial for the laws of motion is the possibility of measuring the time of ob-
servable events by motion. This is not possible for real time. What really
matters in Newtonian physics is thus relative time. Absolute time, dear as it
may have seemed to Newton, is something that lurks in the background,
and is perhaps completely superfluous to the physicist.3
3
His great opponent Gottfried Wilhelm Leibniz argued that time and space are
purely relational – there is no absolute time and no absolute space. In response to
this, Newton’s spokesman Samuel Clarke gave an argument as to why we need
something like “empty space”, independent of the properties of objects that are
“in space”. But no corresponding argument was ever given for “empty time”
(see Westfall 1983).
Concepts of time 11
(a) What is crucial, is not so much the “earlier – later” order of observable
phenomena – their succession; it is their duration. The famous and per-
plexing “time dilation” and “time contraction” effects of the special
theory of relativity refer in the first place to changes in the duration of
some observable phenomena, when measured from different frames of
reference. But indirectly, varying duration also affects observed simul-
taneity and succession between two events. The reason is that informa-
tion about these events needs some time to reach the observer, and this
12 Wolfgang Klein
Since the laws of physics do not conform to an “arrow of time”, which in-
variably flies from earlier to later, they reveal the kind of symmetry which
physicists like; the theory of special relativity started as an attempt to over-
come an undesirable asymmetry. But has nature really no earlier-later ori-
entation? Is the real world, whose laws the physicists try to find, like that?
Questions of this sort have given rise do a different notion of physical time.4
E. We can imagine that an egg, once fried, returns to its initial state; we
can even have a film run backwards, thus apparently reversing the order of
time. But we never observe such a return in reality. There are many physical
processes which, it seems, obey the “arrow of time”. A well-known type of
such unidirectional processes are the changes of entropy (roughly: the
amount of disorder) in a closed system, as studied in thermodynamics. In
Clausius’ formulation from 1851, the second Law of Thermodynamics states
that the overall entropy (roughly: the amount of disorder) of a closed physi-
cal system can remain constant or it can increase; but it cannot decrease,
unless such a change is caused by influences from outside the system: in-
herent state changes of the entire system are unidirectional. This has given
rise to a physical concept of irreversible time, a concept which is neither
Newtonian nor Einsteinian (see, for example, Prigogine and Stengers 1993).
It should be noted, though, that irrreversibility is not to be equated with the
earlier-later asymmetry, as is often done. Even if the fried egg could be re-
4
Reichenbach (1958) is still a very clear treatment of this problem; see also Hor-
wich (1987) and the contributions in Savitt (1993) for a more recent discussion.
Concepts of time 13
stored, the time at which at which it has its original shape again is later
than the time at which it was not yet fried: the egg is as it was before. We
will come back to this problem in section 3.1.
The time of physics, in whichever of the three variants mentioned here, does
not integrate some of the features which we normally associate with time. It
deals with the temporal structure of dead matter, not of living organisms.
There are at least three notions of biological time – the life span of the indi-
vidual, biological evolution, and biological rhythms in the organism.
G. Antique and mediaeval thought did not consider the world as entirely
static. There are changes, such as the motion of bodies, the changing sea-
sons, or even the notion of subsequent ages – for example, a “Golden Age”
followed by a “Silver Age”. But it was not until the late 18th century that
the idea of evolution gained ground – that is, of a temporally directed and
rule-governed process which determines directed changes of whole sys-
tems, usually towards an increasing complexification. The earliest detailed
treatment I am aware of is by Johann Gottfried Herder (1784, vol. I). Such
systems might be, for example, languages; Hermann Paul, one of the lead-
ing linguists of the 19th century, even argued that only historical linguistics
deserves the name of a science, because only this way of looking at lan-
guage reveals the principles that underly it, rather than merely stating facts
Paul 1882: 20). They might be physical systems, such as the earth, the solar
system, or even the entire universe. But by far the most discussed example
is the origin and evolution of life, which, as is now generally believed in the
educated world, is determined by a few principles such as genetic variation,
extinction according to fitness, or drift.
I. What is the Now that separates the past from the future and allows us to
define what is present? This notion has vexed philosophers from Aristotle
to McTaggart, for at least two reasons. First, it “shifts” permanently: there
is not a single now, there are infinitely many nows. But there is always a
special now – the now right now, so to speak. So, how is this now defined
in contrast to all the other nows? Second, the “now” is supposed to have no
extension, hence no duration (and in the sense of physical time, it is not in
time at all: no duration, no time). If this is true, then there can be no pre-
sent. But if there is no present, it seems to make little sense to speak of past
and future. Arguments of this sort have lead to the idea that time is “not
real”, a position indeed taken by philosophers from the antiquity until to
McTaggart. Now, rather than worrying about these puzzles, psychologists
have tried to determine what the minimal unit of perception is, that is, the
shortest time at which our sensory organs can, for example, distinguish a
change in vision or audition. For human beings,5 this shortest moment is
assumed to be somewhere between 30 and 40 milliseconds (Poeppel 1988).
But already William Stern noted in 1897 that this shortest moment may not
coincide with what we consider to be “present” (whose duration he calcu-
5
The idea of such a shortest time span of perception and the possibility that it
might vary across species was first introduced the founding father of embryolo-
gy, Ernst Baer, in 1864. He also beautifully illustrated the dramatic consequences
of this variation for the way in which the world is experienced.
Concepts of time 15
lated as 6–7 seconds). This may not solve the philosophical problems con-
nected to the now; but if the present is defined by what is perceived right
now, then we know at least how long the now is.
N. A second, related aspect is the degree to which daily life and work are
dictated by the mechanical measurement of time. Until a few centuries ago,
Concepts of time 17
O. What role do history and chronology play in a society? All human cul-
tures we know of have some notion of the forebears which may still be
“present” in some sense – dead, but still an active force in daily life. This
connection to the past may be structured in different ways. Some old cul-
tures, such as the Chinese, the Japanese or the Egyptian culture, are bound
to the remote past by an uninterrupted “chain of generations”, for example
dynasties or families. Others, such as the Greek or the Indian culture, also
have strong ties to a very distant past, but they never had the notion of such
a linear chain which connects the present to the origin (see Nakayama
1968).
This brings us to our final point. There are at least four ways in which lan-
guage is crucially connected to time. Languages change in time, they are
processed in time, they exhibit a linear order, and they express time.
Q. In the antiquity and in the middle ages, the idea that languages change
was not unknown; but this fact, obvious as it is, did not play a substantial
role in the way in which philosophers and linguists thought about language.
6
Another interesting case are movies which present a complex event within a cer-
tain time frame, say 90 minutes; but the “real” time of the event thus represented
is, of course, normally much longer. This can be used for special effects, such as
in Buster Keaton’s silent movie “Seven chances” in which movie time and de-
picted time converge, as the movie goes on.
18 Wolfgang Klein
S. There are three major modalities by which human languages are en-
coded – speaking, writing, gesturing. Each utterance, each text follows a
linear order, which is fundamentally temporal in nature. Linguists often
say that a constituent “is moved to the left” or “to the right”; but in fact,
this is only a spatial metaphor for the fact that this constituent is somehow
processed at an earlier or later time when pronounced or written, heard or
read.
In this discussion of reality, two issues must be clearly kept apart: The first
issue concerns the “reality of time”: is time “real”, or is it just a fiction of
our mind. This question has led to vivid discussions, but mainly among phi-
losopers; it is an interesting but somewhat academic problem. The second
issue is the nature of reality itself: is there a reality – maybe the only “real
reality” – behind the apparent changes, which our senses tell us? Different
views which people have taken on this question had dramatic consequences
in the history of mankind; the entire dogma of transsubstantiation, so fun-
damental to Christian faith, depends on the possibility that “real reality” is
independent of apparent change or non-change, and many people have died
for the one position or the other; so, it is probably an important issue.
20 Wolfgang Klein
a) spatial properties, such as being in Crete and in Athens, being here and
there, being under the blanket or on the blanket;
b) qualitative properties, such as being red and green, odd and even, immor-
tal and mortal;
c) quantitative properties, such as getting a bit drunk or heavily drunk, driv-
ing seven miles or 99 miles, weighing one ton or nine tons. These prop-
erties are usually somehow derived, since they operate on qualitative or
spatial changes and indicate differences in degree – either numerically
or in a somewhat fuzzier way.
time. The first confusion concerns the notion of irreversibility, that is, the
“arrow of time” discussed above (section 2, point E). The laws of classical
as well as of relativistic physics apply equally from earlier to later and from
later to earlier: time is “reversible”. Biological time – and similarly the time
of our daily experience – is orientied: it runs from birth to death, never
from death to birth; it is “irreversible”. But this common way to state the
difference is misleading. It is not time that is reversible or irreversible but
the sequence of changes. Take the simple case of a glass which, once bro-
ken, cannot return to the state in which it was not broken (and thus really a
glass, and not a mass of pieces): we can image this, even see it on a film
that runs backwards; but it is never observed in reality. But suppose reality
would indeed allow this to happen – then, there is still an earlier state, in
which the glass is not broken, and a later state, at which it is not broken,
interrupted by a state in-between, at which it is broken: there is a tempo-
rally ordered sequence “not-broken – broken – not broken”, each of which
is associated with a different time: there are three different time spans, two
of which – the first and the last – are associated with the same qualitative
properties. The glass does not “return in time” – it has the same properties
again a later time. So, the difference between “reversible” and “irreversible”
is only whether we can have the sequence of changes “unbroken to broken”
as well as the sequence “broken to unbroken”, or only the former. But each
of these two sequences goes from earlier to later. Even if the order of
changes can go in both directions – time cannot: there is no irreversable
time.
The second confusion concerns the notion of a cyclic time (in contrast to a
linear time). From the Greek notion of the “Great year” – a very long period
after which everything is destroyed by fire and then reborn – to Friedrich
Nietzsche’s “ewige Wiederkehr des Gleichen” (“eternal return of the same”),
many share in the view that the world passes through cycles of creation,
destruction and recreation (the classical treatment is Eliade 1954). In an-
thropology, it is often said that some cultures or some schools of thought do
not have the western notion of linear time: there are time cycles (Wendorff
1980). In linguistics, Benjamin Lee Whorf became famous because of his
claim that the Hopi have a completely different view of time than the one
found in “Standard European Languages”, a view which does not see time
as a linear sequence but as a cycle (see the critical examination in Malotki
1983). But in all of these cases, this does not imply that the time is cyclic. It
only means that the same sequence of changes is repeated and thus cyclic.
The experience of such change cycles is very natural, on a short scale, as the
sequence or day and night, as well as on a larger scale, as the re-appearance
22 Wolfgang Klein
of certain stellar constellations. But this does not mean that the time itself
comes and goes. We can count the repetitions. The seventh time, when the
sun rises in the east, is not the same time as the twelfth time at which this
happens. The twelfth time at which the world is re-created is not the fif-
teenth time at which it is re-created. What might be identical, are the prop-
erties which the world has at these different times.
Aristotle, to whom we owe the first systematic examination of time in
general and of its relation to change in particular, states this very clearly in
his famous definition of time: time is “a number of motion with respect to
before and after” (Physics IV, 219 b 1–2). Aristotle’s analysis of time is not
easy to follow, and it has given rise to various interpretations (see, e.g.,
Coope 2005). But he not only makes a clear distinction between time and
change; without such a distinction, it it would make no sense to say that
something happens slowly or fast. He also characterises time as something
that can be counted. If the entire world is reborn for the seventh time, then
this seventh time is later that the sixth time at which it is reborn.
Is there one time, or are there many times? The common idea is that there is
one time which can be subdivided into smaller units. These “smaller times”
are called time spans, temporal intervals, subtimes, or just “times”. We say
that there is a time at which we met our first love, and a time at which we
lost her or him, and each of these times it is a subinterval of the entire time;
these subintervals themselves have subintervals: time is somehow nested.
Several things can be said about these time spans:
1. They have a duration. We do not know whether the “entire time” has a
duration. As we have seen above (section 2, point D) Newton equated abso-
lute time with duration; but as soon as he talks about the measurement of
time, only smaller time spans – for example the duration of some event –
are at issue.
2. This duration can be measured. This is done by relating the time span,
which is to be measured, to the duration of some other time spans; these are
given by repeated occurrences of specific events (such as the rotation of the
earth). We say that time can be measured. In a way, this is puzzling. When
can it be measured? Clearly, time does not stand still during the measure-
ment process – thus, the entity to be measured changes during this process.
But there is no real puzzle: we do not really measure time, we measure the
Concepts of time 23
duration of events, and we say that this duration is the time during which
the event lasts. But one thing is the event, another thing is its time, just as
there is a difference between a cup and the space which the cup occupies.
3. Time spans do not stand alone. They are related to each other according
to an underlying earlier-later order which is unidirectional. But two time
spans can also be simultaneous or overlap. In other words, time is a sort of
structure whose units are time spans and whose structure is defined by tem-
poral relations such as succession, overlap, simultaneity.
4. Each time span in turn consists of time spans. Does this go on forever –
i.e., is there is “shortest time”? This is surely the case for human time expe-
rience. It is less clear whether nature has a minimal time span. Traditional
as well as modern physicists assumed that natura non fecit saltus (“Nature
does not make jumps”, i.e., there is continuity). It was only in 1900 that
Max Planck showed, quite reluctantly, that physicists are well advised to
assume that there is a shortest time, whose duration is 5.4 x 10-43 seconds.
We can, of course, imagine a shorter time, for example, 10-44 seconds; such
a product or our mind is just meaningless for the laws of physics – and it
would still leave us far away from a continuous time, which has no shortest
interval.
5. Is there is a “last time span”, i.e., does time have an end? And similarly,
is there a “first time span”, i.e., does time have a beginning? St. Augustine
says no, Stephen Hawking says yes, Immanuel Kant says that both views
lead to paradoxes. The reasonable person has no opinion on this issue.
Neither physical time nor biological time, in the senses mentioned in section
2, know the distinction between past, present, and future – notions which
everybody feels to be fundamental to human time. Einstein, in a conversa-
tion with Rudolf Carnap (around 1953), explicitly noted this fact: “Once
Einstein said that the problem of the Now worried him seriously. He ex-
plained that the experience of the Now means something special for man,
something essentially different from the past and the future, but that this
important difference does not and cannot occur within physics. That this
experience cannot be grasped by science seemed to him a matter of painful
but inevitable resignation. I remarked that all that occurs objectively can be
described in science; on the one hand the temporal sequence of events is
described in physics; and, on the other hand, the peculiarities of man’s ex-
24 Wolfgang Klein
periences with respect to time, including his different attitude towards past,
present, and future, can be described and (in principle) explained in psy-
chology. But Einstein thought that these scientific descriptions cannot pos-
sibly satisfy our human needs; that there is something essential about the
Now which is just outside the realm of science. “ (Carnap 1963: 37f.). This
distinction between past, present and future requires an observer; this ob-
server cannot be an instrument which measures time, such as a clock. No
chronometer, precise as it may be, distinguishes past from future. To this
end, an observer is needed who identifies a time span as “being now”. Hu-
man beings are able to do that. Maybe other animals are able to do it as
well, although this question is not easy to answer.
But what is the “now”? In the long philosophical debate on this ques-
tion, there has never been an answer on which the experts agree. Essentially,
there are two different though interconnected problems. First, there is not
just one “now” but infinitely many – the “now” right now, the “nows” that
before that, and the “nows” that are ahead of us. In other words, time itself
seems to be a series of nows. Acccordingly, there is a past and a future rela-
tive to each of these “nows”. But what distinguisthes the “now” right now
from all other “nows”? It must be a special property which somehow comes
from the particular observer who experiences the – inner or outer – world
as somehow “present”, whereas earlier nows are somehow in memory, and
later nows somehow in imagination. But on the other hand, the earlier
“nows” are also defined in relation to the experiences of some observer,
perhaps the same observer at some earlier time; so, the problem cannot be
easily reduced to the difference between memory, experience and expecta-
tion in our soul, as St. Augustine does. It appears, therefore, that the distinc-
tion between “now” and “not-now” is not reducible to any other difference.
The second classical problem results from the fact that the now does not
seem to have an extension; otherwise, it would consist of several moments,
some of which are earlier and hence past, and hence not now. But if the
now has no extension, it does not exist, and hence, there is no presence; but
if there is no presence, there is neither past nor future. Moreover, it is not
possible that the entire time is built up from a series of nows, because if
they have no extension, time cannot have an extension, either. These are
the type of mind-boggling problems that were extensively discussed from
Aristotle (Turetzky 1999: 22–25) to our days (see, for example, Dummett
2000).
In the second half of the 19th century, physiologists and psychologists
set out investigate the notion of “present moment” with experimental meth-
ods. From film watching, everybody knows that when the number of pic-
Concepts of time 25
tures presented to our visual system exceeds about 20 per second, it cannot
keep them apart and perceives them as a continous movement. So, there is a
shortest time for (in this case visual) experience. But does this shortest time
correspond to the “now” which underlies the distinction between past, pre-
sent and future? When watching a film, or listenening to a tune, our intui-
tion about what is on-going, rather than gone or only to come, seems much
longer. So, there must be something in our brain which integrates shortest
perceptual moments into a whole – a “perceptual present”; assumptions go
that this perceptual present can last a few seconds.
Still a different issue is the “now” which underlies the linguistic expres-
sion of past, present and future. All languages in the world mark such a dis-
tinction by tense marking (he is here vs he was here vs he will be here) or
by adverbials such as yesterday, last year, very soon. How is this now de-
fined? Clearly, it cannot be the meaning of a word such as now (or its coun-
terparts in other languages). These words refer to a time span with, as the
case may be, considerable extension (As a child, I was very religious, but
now, I am not). The word now, when uttered in a speech situation, refers to
a time span which INCLUDES the moment of speech, rather to the moment
of speech itself; the boundaries of this time span can vary. It seems to be
this moment of speech which serves as an anchoring point, in relation to
which present, past and future are defined. In fact, this picture is too simple
again because the “moment of speech” is usually not a moment – surely not
in the sense of the shortest time our brain can experience. We shall return to
this problem in section 4. Two facts should be noted, however. First, this
temporal anchoring to whatever the the “present moment” is usually con-
sidered to be fundamental to the expression of time in natural languages.
Second, the temporal anchoring point may differ considerably from what in
other disciplines is considered as “now”.
As we have seen in the preceding two sections, there are many notions of
time, such as biological time, Newton’s absolute and relative time, time as
Kantian “Form der inneren Anschauung” and hence a necessary precondi-
tion of all cognition, subjective time, as influenced for example by drugs,
and so on. These notions are interrelated in many ways, but they cannot be
reduced to one concept: there are many. Is there a concept of time which
underlies the expression of temporal relations in NATURAL LANGUAGES?
Even this is doubtful. In most modern cultures, metrical calendar time plays
26 Wolfgang Klein
an important role, so important that we often take it for self-evident. Our life
is largely organised around (or rather along) this time, and hence, there are
many expressions which refer to it – like in the year of 2007, two hours and
thirty five minutes after noon on May 8, 1998, and so on. But many cultures
do not have such a concept of metrical time, nor the notion of one major
event in collective history to which everything can be temporally related.
Even in Western culture, the full elaboration of this system is fairly recent.
The mere fact that people talk of “hours”, “days”, “years” and “the birth of
Christ” does not mean that they have a concept of metrical time, with the
birth of the saviour, or some other important event, as point zero. Until a
few centuries ago, the concept of “hour”, for example, just meant “twelfth
part of the day”, and if the day was short, like in winter, the hour was short,
as well. A “day” is simply the time, when there is light, or the time from
when people get up until they go to bed again, no matter how “long” this
may be in terms of a mechanical or electronical clock.
Therefore, it seems reasonable to distinguish between various layers of
time structure that are used in the encoding of time. There is something like
a “basic time structure” on which the expression of temporal relations in
natural languages is based. This basic time structure must cover basic rela-
tions between time spans, such as succession and simultaneity, but also the
notion of a basic vantage point – the “now” of an observer. More differen-
tiated structures, like calendaric metrical time, may be added, as cultures
develop. It seems likely, although this is an empirical question, that such
additional structuring is only expressed by more or less complex lexical
expressions, whereas the basic time structure is most often expressed by
grammatical categories and by simple adverbs.7
What, then, is this “basic time structure”? This is not easy to say, because
at most 5% of the world’s languages are sufficiently well described; for all
others, our information is very superficial and often based on bold compari-
sons with familiar languages such as Greek, Latin or English. Hence, we
might simply miss important temporal notions encoded in same or even
many languages. But such is the state of our knowledge. An inspection of
7
The following discussion essentially follows Klein (1984), Chapter 4.
Concepts of time 27
those languages for which our information is more profound shows that the
following six characteristics are indispensable:
B. Inclusion: If s1 and s2 are time spans, then s1 may be included in s2; this
inclusion may be full or partial; in the latter case, we may speak of
“overlapping”.
C. Succession: If s1 and s2 are time spans, which are not (fully or partly)
included in each other, then either s1 precedes s2 or s2 precedes s1.
It is usually said that time is linearly ordered. The way in which we have
characterised succession here is somewhat weaker: there is a partial order on
time spans: time spans can overlap. Again, it is an open question whether
this partial order is based on some full order on “time points”, which make
up the time spans. We normally assume that there is some temporal pro-
gression within a time span, and a strict order on time points allows us to
reconstruct this intuition in a straightforward way.
These three features allow a clear definition of the “earlier-later” asym-
metry between time spans as well as simultaneity. Simultaneity can be full
(two time spans completely coincide) or partial, if they partly overlap.
E. Origo: There is a distinguished time span, which we may call “the time
of present experience”. Everything before that is accessible to us only
by memory, everything later only by expectation.
This origo is the dividing point between past, present and future. As was
discussed in sections 2. (points H. and I.), and 3.3, such an origo is not part
of all time concepts; it plays no role in physical time or in biological time.
But it plays an eminent role in the linguistic encoding of temporal relations.
The best-known case is the grammatical category of tense; in its classical
understanding, tense situates some event in relation to the “deictic origo”,
which is given by the moment of speech – the linguistic variant of the time
of present experience. But there are also many adverbials which are an-
chored at the deictic origo, for example today, three days ago or, of course,
the word now itself; thus, today means “the day which includes the deictic
origo”. Remember, however, that the meaning of the word now is not to be
equated with the deictic origo – it refers to a time span which contains the
deictic origo, but can be much longer (cf. section 3.3). We can say, for ex-
ample Now, the average temperature is much colder than in the pleistocene.
F. Proximity: If s1 and s2 are time spans, then s1 may be near to, or far
from, s2.
This feature is much less discussed in the tradition than linear order, dura-
tion, or the existence of a “now”. But it is regularly encoded in natural lan-
guages. Proximity and non-proximity in this (non-metrical) sense is exem-
plified, for example, by expressions like soon or just; it also sometimes
shows up in tense distinctions, like “near future” vs “far future”. Note that
this concept of “temporal distance” or “remoteness” does not presuppose a
concept of metrical time. Quite the opposite, it is not easy to capture the
idea of proximity in this sense by metrical distance: soon can mean “in ten
minutes”, like the meal will be served soon; but it can also mean “ten
months”, like in they soon got divorced again.
G. Lack of quality: Time spans have no qualitative properties; they are nei-
ther green nor sweet, and they have no wheels and no spines. They are
contained in each other or just after each other or or more or less close
to each other, and they are long or short.
In the tradition, this feature shows up in the discussion about how time and
change are related to each other (see section 3.1 above). The latter normally
Concepts of time 29
(a) The most straightforward way to account for the notion of “origo” is to
identify it with the moment of speech; in fact, such a “deictic origo” is
found in all human languages we know.
(b) It is less clear how one should capture our intuitive notion of (tempo-
rally) nearness. One might think to use the natural topology on the real
numbers: the neighbourhoods of any real number r are exactly those
open intervals to which r belongs. But this gives us by far too much: it
gives us all environments, rather than the one which marks the border-
line between “close” and “far”. Our intuitive notions tell us that each
time span has a “REGION” around itself, whose borders vary with con-
text. The time of drinking a coffee is usually shorter than the time of
finding a proof, and so are the “regions” around these two time spans.
Temporal relations between two time spans s and t do not only differ
30 Wolfgang Klein
Temporal relations obtain between two time spans: a first time span, which
I will call THEME, and some other time span, which I will call RELATUM.
The what follows, theme is marked by – – – – – – , the relatum is marked by
+ + + + +, and the region around the relatum by ( ); the linear order is repre-
sented by left-right arrangement:
b. LONG BEFORE, i.e,, the theme precedes the region of the relatum:
– – – – – – ( +++++ )
c. SHORTLY BEFORE, i.e., the theme precedes the relatum, but it is in the
region of the relatum:
( – – – – – – + + + ++ )
d. JUST BEFORE, i.e., as SHORTLY BEFORE, but the theme abuts the
relatum:
– – – – – – +++++
8
In all of these cases, the Basic Time Structure allows us to give precise formal
definitions. For present purposes, however, it will be more useful to use diagrams
that illustrate the various relations.
32 Wolfgang Klein
e. PARTLY BEFORE, i.e., a the first part of the theme precedes and the
second part of the theme is IN the relatum (the region is irrelevant):
– – –+–+–+–++
Calendaric relata is of lesser interest here; they only differ in which histori-
cal event from the shared knowledge of the interlocutors is chosen as an
anchoring point – the foundation of Rome, birth of Christ, the Hedjra, the
beginning of a dynasty, etc. There is no language in which tense is linked to
Concepts of time 33
We also have the opposite problem, i.e., cases in which the “time of utter-
ance” seems to go beyond the boundaries of a single sentence. Does a
longer text, say a lecture or even a novel, have a single time of utterance or
a different one for each single utterance of which it consists? In a sense, a
coherent sequence of utterances – a text, be it written or spoken – is a unit,
and it should have a single relatum. But then, it would be strange to assume
that this relatum is, for example, the time at which the whole text was pro-
duced: What is then the utterance time of the Bible, or the first book of
Moses? In these cases, the characterisation of the deictic relatum as “time
of utterance” is clearly insufficient.
Linguistic systems most often evolve in spoken communication, in which
speaker and listener are equally present. Then, time of speaking and time of
34 Wolfgang Klein
In these cases, it is not the origo of the real speaker which counts but the
origo of the person whose thinking or speaking is being talked about.
Anaphoric relata are time spans which are given somewhere in the linguis-
tic context. Their role for tense is disputed. In the literature, a distinction is
often made between “absolute” and “relative” tenses; the former are purely
deictic, whereas the latter also involve an anaphoric relatum. Some text
types, for example narratives, are based on an chain of such anaphoric re-
lata. As with all types of anaphora, there are three subcases:
2. The anaphoric relation may go from one clause to another, whilst still
being in the same sentence (interclausal anaphora):
(6) When the phone rang, he switched the light off.
Concepts of time 35
In cases of this type, it is often said that “two events” are temporally related
to each other. But note that the entire when-clause only serves to define a
time span, which functions as a relatum. In principle, this is not different
from the anaphoric relatum in (5), which is simply specified by a clock-
time adverbial.
5. Concluding remarks
Acknowledgement
References
Flasch, Kurt
2004 Was ist Zeit?. Augustinus von Hippo, das XI. Buch der Confessiones.
Frankfurt a. M.: Klostermann.
Foster, Russell G. and Leon Kreitzman
2004 Rhythms of life: The biological clocks that control the daily lives of
every living thing. London: Profile Books.
Fraser, Julius T.
1987 Time – the familiar stranger. Amherst: University of Massachusetts
Press.
Fraser, Julius T. (ed.)
1968 The voices of time. London: Penguin
Fraser, Julius T., Francis C. Haber and Nathaniel M. Lawrence (eds.)
1986 Time, Science, and Society in China and the West. Amherst: Univer-
sity of Massachusetts Press.
Heidegger, Martin
1927 Sein und Zeit. Jena: Niemeyer.
Herder, Johann Gottfried
1784 Ideen zu einer Philosophie der Geschichte der Menschheit. Riga:
Hartknoch.
Horwich, Paul
1987 Asymmetries in Time. Cambridge, MA: MIT Press.
Janda, Richard D. and Brian D. Joseph (eds.)
2004 The Handbook of Historical Linguistics. Oxford: Blackwell.
Jourdain, Robert
1997 Music, the Brain, and Ecstasy: How Music Captures Our Imagina-
tion. New York: William Morrow.
Kant, Immanuel
1781 Kritik der reinen Vernunft. Riga: Hartknoch.
Klein, Wolfgang
1994 Time in Language. London: Routledge.
Le Poidevin, Robin and Murray McBeath (eds.)
1993 The Philosophy of Time. Oxford: Oxford University Press.
Libet, Benjamin
2004 Mind Time. The Temporal Factor in Consciousnes. Cambridge, MA:
Harvard University Press.
Lucas, John R.
1999 A century of time. In Butterfield 1999: 1–20.
Malotki, Ekkehart
1983 Hopi Time. Berlin /New York: Mouton de Gruyter.
McTaggert, John E. M.
1908 The Unreality of Time. Mind 17: 457– 474. (Reprinted in Le Poidevin
and McBearth 1993).
38 Wolfgang Klein
Nakayama, Hajime
1968 Time in Indian and Japanese thought. In Fraser 1968: 77–91.
Newton, Isaac
1972 Isaac Newton’s Philosophiae naturalis principia mathematica. The
3rd edition, with variant readings, assembled and edited by Alexandre
Koyré and Bernard Cohen. Cambridge, MA: Harvard University Press.
Paul, Hermann
1882 Prinzipien der Sprachgeschichte. 2nd Edition. Jena: Niemeyer.
Piaget, Jean
1923 Le développement de la notion de temps chez l’enfant. Paris: Presses
Universitaires de France.
Poeppel, Ernst
1988 Mindworks: Time and Conscious Experience. Boston: Harcourt.
Prigogine, Ilya, and Stengers, Isabelle
1993 Time, Chaos and the Quantum: Towards the Resolution of the Time
Paradox. New York: Harmony Books.
Reichenbach, Hans
1958 The Direction of Time. Berkeley, CA: University of California Press.
Richards, Edward G.
1998 Mapping Time, the calendar and its history. Oxford: Oxford Univer-
sity Press.
Savitt, Steven, ed.
1995 Time’s Arrows Today: Recent Physical and Philosophical Work on
the Direction of Time. Cambridge: Cambridge University Press.
Turetzky, Philip
1998 Time. London: Routledge.
Whitrow, Geoffrey J.
1980 The natural philosophy of time. Oxford: Clarendon Press.
Wendorff, Rudolf
1980 Zeit und Kultur. Geschichte des Zeitbewußtseins in Europa. Wies-
baden: Westdeutscher Verlag.
Westfall, Richard Samual
1983 Never at rest. A biography of Isaac Newton. Cambridge: Cambridge
University Press.