![Chapter 1
A Conceptual Framework for Developmental
Cognitive Systems
1.1 Introduction
This book addresses the central role played by development in cognition. We are in-
terested in particular in applying our knowledge of development in natural cognitive
systems, i.e. human infants, to the problem of creating artificial cognitive systems in
the guise of humanoid robots. Thus, our subject matter is cognition, development,
and humanoid robotics. These three threads are woven together to form a roadmap
that when followed will enable the instantiation and development of an artificial
cognitive system. However, to begin with, we must be clear what we mean by the
term cognition so that, in turn, we can explain the pivotal role of development and
the central relevance of humanoid embodiment.
In the following, we present a conceptual framework that identifies and explains
the broad stance we take on cognitive systems — emergent embodied systems that
develop cognitive skills as a result of their action in the world — and that draws out
explicitly the theoretical and practical consequences of adopting this stance.1
We begin by considering the operational characteristics of a cognitive system,
focussing on the purpose of cognition rather than debating the relative merits of
competing paradigms of cognition. Of course, such a debate is important because
it allows us to understand the pre-conditions for cognition so, once we have estab-
lished the role cognition plays and see why it is important, we move on to elaborate
on these pre-conditions. In particular, we introduce the underlying framework of
enaction which we adopt as the basis for the research described in this book.
By working through the implications of the enactive approach to cognition, the
central role of development in cognition becomes clear, as do several other key
issues such as the crucial role played by action, the inter-dependence of percep-
tion and action, and the consequent constructivist nature of the cognitive system’s
knowledge.
1 This chapter is based directly on a study of enaction as a framework for development in
cognitive robotics [385]. The original paper contains additional technical details relating
to enactive systems which are not strictly required here. Readers who are interested in
delving more deeply into enaction are encouraged to refer to the original.
D. Vernon et al.: A Roadmap for Cogn. Develop. in Humanoid Robots, COSMOS 11, pp. 1–11.
springerlink.com c
Springer-Verlag Berlin Heidelberg 2010](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-15-320.jpg)
![2 1 A Conceptual Framework for Developmental Cognitive Systems
The framework of enaction provides the foundation for subsequent chapters
which deal with the phylogeny and the ontogeny of natural cognitive systems —
their initial capabilities and subsequent development — and the application of what
we learn from these to the realization of an artificial cognitive system in the form of
a humanoid robot.
1.2 Cognition
Cognitive systems anticipate, assimilate, and adapt. In doing so, they learn and de-
velop [387]. Cognitive systems anticipate future events when selecting actions, they
subsequently learn from what actually happens when they do act, and thereby they
modify subsequent expectations and, in the process, they change how the world
is perceived and what actions are possible. Cognitive systems do all of this au-
tonomously. The adaptive, anticipatory, autonomous viewpoint reflects the position
of Freeman and Núñez who, in their book Reclaiming Cognition [105], assert the
primacy of action, intention, and emotion in cognition. In the past, however, cog-
nition was viewed in a very different light as a symbol-processing module of mind
concerned with rational planning and reasoning. Today, however, this is changing
and even proponents of these early approaches now see a much tighter relationship
between perception, action, and cognition (e.g. see [7, 214]).
So, if cognitive systems anticipate, assimilate, and adapt, if they develop and
learn, the first question to ask is why do they do this? The subsequent question is
how do they do it? The remainder of this section is devoted to the first question and
the rest of the book addresses the latter.
The view of cognition taken in this book is that cognition is the process whereby
an autonomous self-governing system acts effectively in the world in which it is
embedded [237]. However, in natural systems, the latencies inherent in the neu-
ral processing of sense data are too great to allow effective action. This is one of
the primary reasons a cognitive agent must anticipate future events: so that it can
prepare the actions it may need to take. In addition, there are also limitations im-
posed by the environment and the cognitive system’s body. To perform an action,
one needs to have the relevant body part in a certain place at a certain time. In a
dynamic environment that is constantly changing and with a body that takes time to
move, this requires preparation and prediction. Furthermore, the world in which the
agent is embedded is unconstrained and the sensory data which is available to the
cognitive system is not only ‘out-of-date’ but it is also uncertain and incomplete.
Consequently, it is not possible to encapsulate a priori all the knowledge required
to deal successfully with the circumstances it will experience so that it must also
be able to adapt, progressively increasing its space of possible actions as well as
the time horizon of its prospective capabilities. It must do this, not as a reaction
to external stimuli but as a self-generated process of proactive understanding. This
process is what we mean by development. In summary, and as noted in the Preface,
the position being set out in this book is that (a) cognition is the process by which
an autonomous self-governing agent acts effectively in the world in which it is em-
bedded, that (b) the dual purpose of cognition is to increase the agent’s repertoire](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-16-320.jpg)
![1.3 Enaction 3
of effective actions and its power to anticipate the need for and outcome of future
actions, and that (c) development plays an essential role in the realization of these
cognitive capabilities.
We will now introduce a framework which encapsulates all these considerations.
1.3 Enaction
There are many alternative perspectives on cognition: what it is, why it is necessary,
and how it is achieved. We have already argued that cognition arises from an agent’s
need to compensate for latencies in neural processing by anticipating what may be
about to happen and by preparing its actions accordingly. So we can agree fairly
easily what cognition is — a process of anticipating events and acting appropriately
and effectively — and why it is necessary — to overcome the physical limitations
of biological brains and the limitations of bodily movements operating in a dynamic
environment. The difficulty arises when we consider how cognition is achieved.
There are several competing theories of cognition, each of which makes its own
set of assumptions. Here, we wish to focus on one particularly important paradigm
— enaction — and pick out its most salient aspects in order to provide a sound
theoretical foundation for the role of development in cognition [235, 236, 237, 238,
359, 380, 382, 381, 400] .
The principal idea of enaction is that a cognitive system develops it own under-
standing of the world around it through its interactions with the environment. Thus,
enaction entails that the cognitive system operates autonomously and that it gener-
ates its own models of how the world works. When dealing with enactive systems,
there are five key elements to consider [280, 373]:
1. Autonomy
2. Embodiment
3. Emergence
4. Experience
5. Sense-making
Autonomy is the self-maintaining organizational characteristic of living creatures
that enables them to use their own capacities to manage their interactions with the
world, and with themselves, in order to remain viable [61, 108]. This simply means
that the system is entirely self-governing and self-regulating: it is not controlled by
any outside agency and this allows it to stand apart from the rest of the environment
and operate independently of it. That’s not to say that the system isn’t influenced
by the world around it, but rather that these influences are brought about through
interactions that don’t threaten the autonomous operation of the system.
The second element of enaction is the idea of embodiment. For our purposes
here, embodiment means that the system must exist in the world as a physical entity
which can interact directly with the environment. This means the system can act on
things in the world around it and they, in turn, can act on the system. These things
can be inanimate objects or animate agents, cognitive or not. As it happens, there](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-17-320.jpg)
![1.3 Enaction 5
Fig. 1.1 Maturana and Varela’s ideograms to denote structurally-determined autopoietic and
organizationally-closed systems. The arrow circle denotes the autonomy, self-organization,
and self-production of the system, the rippled line the environment, and the bi-directional
half-lines the mutual perturbation — structural coupling — between the two.
sense-making and the resultant knowledge says nothing at all about what is really
out there in the environment; it doesn’t have to. All it has to do is make sense for the
continued existence and autonomy of the cognitive system. Sense-making is actu-
ally the source of the term enaction. In making sense of its experience, the cognitive
system is somehow bringing out through its actions — enacting — what is important
for the continued existence of the system. This enaction is effected by the system
as it is embedded in its environment, but as an autonomous entity distinct from the
environment, through an emergent process of making sense of its experience. This
sense-making is, in fact, cognition [108].
The founders of the enactive approach, Maturana and Varela, introduced a dia-
grammatic way of conveying the self-organizing and self-maintaining autonomous
nature of an enactive system, perturbing and being perturbed by its environment
[237]: see Fig. 1.1. The arrowed circle denotes the autonomy and self-organization
of the system, the rippled line the environment, and the bi-directional half-arrows
the mutual perturbation.
1.3.1 Enaction and Development
So what has all this to do with development? Our position in this book is that learn-
ing arises as a consequence of the interaction between the cognitive agent and the
world around it, whereas development arises from learning through a process of in-
teraction of the cognitive agent with itself. We remarked above that the process of
sense-making forms a virtuous circle in that the self-generated model of the sys-
tem’s experience provides a larger repertoire of possible actions, richer interactions,
increased perceptual capacity, and potentially better self-generated models, and so
on. Recall also our earlier remarks that the cognitive system’s knowledge is rep-
resented by the state of the system. When this state is embodied in the system’s
central nervous system, the system has much greater plasticity in two senses: (a) the
nervous system can accommodate a much larger space of possible associations be-
tween system-environment interactions, and (b) it can accommodate a much larger
space of potential actions. Consequently, the process of cognition involves the sys-
tem modifying its own state, specifically its central nervous system, as it enhances](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-19-320.jpg)
![6 1 A Conceptual Framework for Developmental Cognitive Systems
its predictive capacity and its action capabilities. This is exactly what we mean by
development. This generative (i.e. self-constructed) autonomous learning and devel-
opment is one of the hallmarks of the enactive approach [280, 108].
Development, then, is identically the cognitive process of establishing and en-
larging the possible space of mutually-consistent couplings in which a system can
engage (or, perhaps more appropriately, which it can withstand without compro-
mising its autonomy). The space of perceptual possibilities is predicated not on an
absolute objective environment, but on the space of possible actions that the system
can engage in whilst still maintaining the consistency of the coupling with the en-
vironment. These environmental perturbations don’t control the system since they
are not components of the system (and, by definition, don’t play a part in the self-
organization) but they do play a part in the ontogenetic development of the system.
Through this ontogenetic development, the cognitive system develops its own epis-
temology, i.e. its own system-specific history- and context-dependent knowledge
of its world, knowledge that has meaning exactly because it captures the consis-
tency and invariance that emerges from the dynamic self-organization in the face
of environmental coupling. Again, it comes down to the preservation of auton-
omy, but this time doing so in an every increasing space of autonomy-preserving
couplings.
This process of development is achieved through self-modification by virtue of
the presence of a central nervous system: not only does environment perturb the
system (and vice versa) but the system also perturbs itself and the central nervous
system adapts as a result. Consequently, the system can develop to accommodate a
much larger space of effective system action. This is captured in a second ideogram
of Maturana and Varela (see Fig. 1.2) which adds a second arrow circle to the
ideogram to depict the process of development through self-perturbation and self-
modification. In essence, development is autonomous self-modification and requires
the existence of a viable phylogeny, including a nervous system, and a suitable
ontogeny.
Fig. 1.2 Maturana and Varela’s ideograms to denote structurally-determined autopoietic and
operationally-closed systems. The diagram (denotes an organizationally-closed autonomous
system with a central nervous system. This system is capable of development by means of
self-modification of its nervous system, so that it can accommodate a much larger space of
effective system action.](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-20-320.jpg)
![1.3 Enaction 7
1.3.2 Enaction and Knowledge
Let us now move on to discuss in a little more detail the nature of the knowledge that
an enactive cognitive system constructs. This knowledge is built on sensorimotor as-
sociations, achieved initially by exploration of what the world offers. However, this
is only the beginning. The enactive system uses the knowledge gained to form new
knowledge which is then subjected to empirical validation to see whether or not it is
warranted (we, as enactive beings, imagine many things but not everything we imag-
ine is plausible or corresponds well with reality). One of the key issues in cognition
is the importance of internal simulation in accelerating the scaffolding of this early
developmentally-acquired sensorimotor knowledge to provide a means to predict
future events, to reconstruct or explain observed events (constructing a causal chain
leading to that event), or to imagine new events [132, 144, 337]. Naturally, there is
a need to focus on (re-)grounding predicted, explained, or imagined events in expe-
rience so that the system can do something new and interact with the environment
in a new way. If the cognitive system wishes or needs to share this knowledge with
other cognitive systems or communicate with other cognitive systems, it will only
be possible if they have shared a common history of experiences and if they have a
similar phylogeny and a compatible ontogeny.
In other words, the meaning of the knowledge that is shared is negotiated and
agreed by consensus through interaction.
When there are two or more cognitive agents involved, interaction is a shared
activity in which the actions of each agent influence the actions of the other agents
engaged in the same interaction, resulting in a mutually constructed pattern of
shared behaviour [275]. Again, Maturana and Varela introduce a succinct diagram-
matic way of of conveying this coupling between cognitive agent and the develop-
ment it engenders [238]: see Fig. 1.3. Such mutually-constructedpatterns of comple-
mentary behaviour is also emphasized in Clark’s notion of joint action [64]. Thus,
explicit meaning is not necessary for anything to be communicated in an interac-
tion, it is simply important that the agents are mutually engaged in a sequence of
actions. Meaning emerges through shared consensual experience mediated by in-
teration. The research programme encapsulated in this roadmap is based on this
foundational principle of interaction. The developmental progress of imitation fol-
lows tightly that of the development of other interactive and communicative skills,
Fig. 1.3 Maturana and Varela’s ideogram to denote the development engendered by interac-
tion between cognitive systems](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-21-320.jpg)
![heat of the sun, and with a shade of banana leaves for her eyes. There,
dripping with sweat in the burning sun and coated with mud to the hips and
over the elbows, she toils to set the younger women a good example.
Moreover, as in every other occupation, the kaliths, the gods, must also be
invoked, and who could be better fitted for the discharge of so important a
duty than the Mother of the House?”522
It seems clear that in any
agricultural people who, like the Pelew Islanders, retain mother-kin and
depute the labours of husbandry to women, the conception of a great
Mother Goddess, the divine source of all fertility, might easily originate.
Perhaps the same social and industrial conditions may have combined to
develop the great Mother Goddesses of Western Asia and Egypt.
But in the Pelew Islands women have yet another road to power. For some
of them are reputed to be the wives of gods, and act as their oracular
mouthpieces. Such prophetesses are called Amlaheys, and no surprise is felt
when one of them is brought to bed. Her child passes for the offspring of
the god, her divine husband, and goes about with his hair hanging loose in
token of his superhuman parentage. It is thought that no mortal man would
dare to intrigue with one of these human wives of a god, since the jealous
deity would surely visit the rash culprit with deadly sickness and a lingering
decline.523
But in these islands men as well as women are often possessed
by a deity and speak in his name. Under his inspiration they mimic, often
with great histrionic skill, the particular appearance and manner which are
believed to be characteristic of the indwelling divinity. These inspired men
(Korongs) usually enjoy great consideration and exert a powerful influence
over the whole community. They always acquire wealth in the exercise of
their profession. When they are not themselves chiefs, they are treated as
chiefs or even preferred to them. In not a few places the deity whom [pg
208] they personate is also the political head of the land; and in that case
his inspired priest, however humble his origin, ranks as a spiritual king and
rules over all the chiefs. Indeed we are told that, with the physical and
intellectual decay of the race, the power of the priests is more and more in
the ascendant and threatens, if unchecked, to develop before long into an
absolute theocracy which will swallow up every other form of government.524
Thus the present, or at least the recent, state of society and religion in the
Pelew Islands presents some interesting parallels to the social and religious
condition of Western Asia and Egypt in early days, if the conclusions
reached in this work are correct. In both regions we see a society based on
mother-kin developing a religion in which goddesses of the clan originally](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-23-320.jpg)
![f
f
f
occupied the foremost place, though in later times, as the clans coalesced
into states, the old goddesses have been rivalled and to some extent
supplanted by the new male gods of the enlarged pantheon. But in the
religion of the Pelew Islanders, as in that of the Khasis and the ancient
Egyptians, the balance of power has never wholly shifted from the female to
the male line, because society has never passed from mother-kin to father-
kin. And in the Pelew Islands as in the ancient East we see the tide of
political power running strongly in the direction of theocracy, the people
resigning the conduct of affairs into the hands of men who claimed to rule
them in the name of the gods. In the Pelew Islands such men might have
developed into divine kings like those of Babylon and Egypt, if the natural
course of evolution had not been cut short by the intervention of Europe.525
The evidence of the Khasis and the Pelew Islanders, two peoples very
remote and very different from each other, suffices to prove that the
influence which mother-kin may exert on religion is real and deep. But in
order [pg 209] to dissipate misapprehensions, which appear to be rife on
this subject, it may be well to remind or inform the reader that the ancient
and widespread custom of tracing descent and inheriting property through
the mother alone does not by any means imply that the government of the
tribes which observe the custom is in the hands of women; in short, it
should always be borne in mind that mother-kin does not mean mother-rule.
On the contrary, the practice of mother-kin prevails most extensively
amongst the lowest savages, with whom woman, instead of being the ruler
of man, is always his drudge and often little better than his slave. Indeed,
so far is the system from implying any social superiority of women that it
probably took its rise from what we should regard as their deepest
degradation, to wit, from a state of society in which the relations of the
sexes were so loose and vague that children could not be fathered on any
particular man.526
When we pass from the purely savage state to that higher plane of culture
in which the accumulation of property, and especially of landed property,
has become a powerful instrument of social and political influence, we
naturally find that wherever the ancient preference for the female line of
descent has been retained, it tends to increase the importance and enhance
the dignity of woman; and her aggrandizement is most marked in princely
families, where she either herself holds royal authority as well as private
property, or at least transmits them both to her consort or her children. But
this social advance of women has never been carried so far as to place men](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-24-320.jpg)
![f
f
as a whole in a position of political subordination to them. Even where the
system of mother-kin in regard to descent and property has prevailed most
fully, the actual government has generally, if not invariably, remained in the
hands of men. Exceptions have no doubt occurred; women have
occasionally arisen [pg 210] who by sheer force of character have swayed
for a time the destinies of their people. But such exceptions are rare and
their effects transitory; they do not affect the truth of the general rule that
human society has been governed in the past and, human nature remaining
the same, is likely to be governed in the future, mainly by masculine force
and masculine intelligence.
To this rule the Khasis, with their elaborate system of mother-kin, form no
exception. For among them, while landed property is both transmitted
through women and held by women alone, political power is transmitted
indeed through women, but is held by men; in other words, the Khasi tribes
are, with a single exception, governed by kings, not by queens. And even in
the one tribe, which is nominally ruled by women, the real power is
delegated by the reigning queen or High Priestess to her son, her nephew,
or a more distant male relation. In all the other tribes the kingship may be
held by a woman only on the failure of all male heirs in the female line.527
So far is mother-kin from implying mother-rule. A Khasi king inherits power
in right of his mother, but he exercises it in his own. Similarly the Pelew
Islanders, in spite of their system of mother-kin, are governed by chiefs, not
by chieftainesses. It is true that there are chieftainesses, and that they
indirectly exercise much influence; but their direct authority is limited to the
affairs of women, especially to the administration of the women's clubs or
associations, which [pg 211] answer to the clubs or associations of the
men.528
And to take another example, the Melanesians, like the Khasis and
the Pelew Islanders, have the system of mother-kin, being similarly divided
into exogamous clans with descent in the female line; “but it must be
understood that the mother is in no way the head of the family. The house
of the family is the father's, the garden is his, the rule and government are
his.”529
We may safely assume that the practice has been the same among all the
many peoples who have retained the ancient system of mother-kin under a
monarchical constitution. In Africa, for example, the chieftainship or
kingship often descends in the female line, but it is men, not women, who
inherit it.530
The theory of a gynaecocracy is in truth a dream of visionaries
and pedants. And equally chimerical is the idea that the predominance of](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-25-320.jpg)
![f
f
goddesses under a system of mother-kin like that of the Khasis is a creation
of the female mind. If women ever created gods, they would be more likely
to give them masculine than feminine features. In point of fact the great
religious ideals which have permanently impressed themselves on the world
seem always to have been a product of the male imagination. Men make
gods and women worship them. The combination of ancestor-worship with
mother-kin furnishes a simple and sufficient explanation of the superiority of
goddesses over gods in a state of society where these conditions prevail.
Men naturally assign the first place in their devotions to the ancestress from
whom they trace their descent. We need not resort to a fantastic hypothesis
of the preponderance of the feminine fancy in order to account for the facts.
The theory that under a system of mother-kin the women rule the men and
set up goddesses for them to [pg 212] worship is indeed so improbable in
itself, and so contrary to experience, that it scarcely deserves the serious
attention which it appears to have received.531
But when we have brushed
aside these cobwebs, as we must do, we are still left face to face with the
solid fact of the wide prevalence of mother-kin, that is, of a social system
which traces descent and transmits property through women and not
through men. That a social system so widely spread and so deeply rooted
should have affected the religion of the peoples who practise it, may
reasonably be inferred, especially when we remember that in primitive
communities the social relations of the gods commonly reflect the social
relations of their worshippers. How the system of mother-kin may mould
religious ideas and customs, creating goddesses and assigning at least a
nominal superiority to priestesses over priests, is shown with perfect lucidity
by the example of the Khasis, and hardly less clearly by the example of the
Pelew Islanders. It cannot therefore be rash to hold that what the system
has certainly done for these peoples, it may well have done for many more.
But unfortunately through lack of documentary evidence we are seldom able
to trace its influence so clearly.](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-26-320.jpg)
![§ 3. Mother-Kin and Mother Goddesses in the Ancient East.
While the combination of mother-kin in society with a preference for
goddesses in religion is to be found as a matter of fact among the Khasis
and Pelew Islanders of to-day, the former prevalence of mother-kin in the
lands where the great goddesses Astarte and Cybele were worshipped is a
matter of inference only. In later times father-kin had certainly displaced
mother-kin among the Semitic worshippers of Astarte, and probably the
same change had taken place among the Phrygian worshippers of Cybele.
Yet the older [pg 213] custom lingered in Lycia down to the historical
period;532
and we may conjecture that in former times it was widely spread
through Asia Minor. The secluded situation and rugged mountains of Lycia
favoured the survival of a native language and of native institutions long
after these had disappeared from the wide plains and fertile valleys which
lay on the highroads of war and commerce. Lycia was to Asia Minor what
the highlands of Wales and of Scotland have been to Britain, the last
entrenchments where the old race stood at bay. And even among the
Semites of antiquity, though father-kin finally prevailed in matters of descent
and property, traces of an older system of mother-kin, with its looser sexual
relations, appear to have long survived in the sphere of religion. At all
events one of the most learned and acute of Semitic scholars adduced what
he regarded as evidence sufficient to prove “that in old Arabian religion gods
and goddesses often occurred in pairs, the goddess being the greater, so
that the god cannot be her Baal, that the goddess is often a mother without
being a wife, and the god her son, and that the progress of things was
towards changing goddesses into gods or lowering them beneath the male
deity.”533
In Egypt the archaic system of mother-kin, with its preference for women
over men in matters of property and inheritance, lasted down to Roman
times, and it was traditionally [pg 214] based on the example of Isis, who
had avenged her husband's murder and had continued to reign after his
decease, conferring benefits on mankind. “For these reasons,” says Diodorus
Siculus, “it was appointed that the queen should enjoy greater power and
honour than the king, and that among private people the wife should rule](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-27-320.jpg)
![over her husband, in the marriage contract the husband agreeing to obey
his wife in all things.”534
A corollary of the superior position thus conceded to
women in Egypt was that the obligation of maintaining parents in their old
age rested on the daughters, not on the sons, of the family.535
The same legal superiority of women over men accounts for the most
remarkable feature in the social system of the ancient Egyptians, to wit, the
marriage of full brothers with full sisters. That marriage, which to us seems
strange and unnatural, was by no means a whim of the reigning Ptolemies;
on the contrary, these Macedonian conquerors appear, with characteristic
prudence, to have borrowed the custom from their Egyptian predecessors
for the express purpose of conciliating native prejudice. In the eyes of the
Egyptians “marriage between brother and sister was the best of marriages,
and it acquired an ineffable degree of sanctity when the brother and sister
who contracted it were themselves born of a brother and sister, who had in
their turn also sprung from a union of the same sort.”536
Nor did the
principle apply only to gods and kings. The common people acted on it in
their daily life. They regarded marriages between brothers and sisters as the
most natural and reasonable of all.537
The evidence of legal documents, [pg
215] including marriage contracts, tends to prove that such unions were the
rule, not the exception, in ancient Egypt, and that they continued to form
the majority of marriages long after the Romans had obtained a firm footing
in the country. As we cannot suppose that Roman influence was used to
promote a custom which must have been abhorrent to Roman instincts, we
may safely assume that the proportion of brother and sister marriages in
Egypt had been still greater in the days when the country was free.538
It would doubtless be a mistake to treat these marriages as a relic of
savagery, as a survival of a tribal communism which knew no bar to the
intercourse of the sexes. For such a theory would not explain why union
with a sister was not only allowed, but preferred to all others. The true
motive of that preference was most probably the wish of brothers to obtain
for their own use the family property, which belonged of right to their
sisters, and which otherwise they would have seen in the enjoyment of
strangers, the husbands of their sisters. This is the system which in Ceylon
is known as beena marriage. Under it the daughter, not the son, is the heir.
She stays at home, and her husband comes and lives with her in the house;
but her brother goes away and dwells in his wife's home, inheriting nothing
from his parents.539
Such a system could not fail in time to prove irksome.
Men would be loth to quit the old home, resign the ancestral property to a](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-28-320.jpg)
![stranger, and go out to seek their fortune empty-handed in the world. The
remedy was obvious. A man had nothing to do but to marry his sister
himself instead of handing her over to another. Having done so he stayed at
home and enjoyed the family estate in virtue of his marriage with the
heiress. This simple and perfectly effective expedient for keeping the
property in the [pg 216] family most probably explains the custom of
brother and sister marriage in Egypt.540
Thus the union of Osiris with his sister Isis was not a freak of the story-
teller's fancy: it reflected a social custom which was itself based on practical
considerations of the most solid kind. When we reflect that this practice of
mother-kin as opposed to father-kin survived down to the latest times of
antiquity, not in an obscure and barbarous tribe, but in a nation whose
immemorial civilization was its glory and the wonder of the world, we may
without being extravagant suppose that a similar practice formerly prevailed
in Syria and Phrygia, and that it accounts for the superiority of the goddess
over the god in the divine partnerships of Adonis and Astarte, of Attis and
Cybele. But the ancient system both of society and of religion had
undergone far more change in these countries than in Egypt, where to the
last the main outlines of the old structure could be traced in the national
institutions to which the Egyptians clung with a passionate, a fanatical
devotion. Mother-kin, the divinity of kings and queens, a sense of the
original connexion of the gods with nature—these things outlived the
Persian, the Macedonian, the Roman conquest, and only perished under the
more powerful solvent of Christianity. But the old order did not vanish at
once with the official establishment of the new religion. In the age of
Constantine the Greeks of Egypt still attributed the rise of the Nile to
Serapis, the later form of Osiris, alleging [pg 217] that the inundation could
not take place if the standard cubit, which was used to measure it, were not
deposited according to custom in the temple of the god. The emperor
ordered the cubit to be transferred to a church; and next year, to the
general surprise, the river rose just as usual.541
Even at a later time
Athanasius himself had to confess with sorrow and indignation that under
his own eyes the Egyptians still annually mourned the death of Osiris.542
The
end came with the destruction of the great Serapeum at Alexandria, the last
stronghold of the heathen in Egypt. It perished in a furious and bloody
sedition, in which Christians and pagans seem to have vied with each other
in mutual atrocities. After its fall the temples were levelled with the ground](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-29-320.jpg)
![f
f
f
or converted into churches, and the images of the old gods went to the
melting-pot to be converted into base uses for the rabble of Alexandria.543
The singular tenacity with which the Egyptian people maintained their
traditional beliefs and customs for thousands of years sprang no doubt from
the stubborn conservatism of the national character. Yet that conservatism
was itself in great measure an effect of geographical and climatic conditions
and of the ways of life which they favoured. Surrounded on every side by
deserts or almost harbourless seas, the Egyptians occupied a position of
great natural strength which for long ages together protected them from
invasion and allowed their native habits to set and harden, undisturbed by
the subversive influence of foreign conquest. The wonderful regularity of
nature in Egypt also conduced to a corresponding stability in the minds of
the people. Year in, year out, the immutable succession of the seasons
brought with it the same unvarying round of agricultural toil. What the
fathers had done, the sons did in the same manner at the same season, and
so it went on from [pg 218] generation to generation. This monotonous
routine is common indeed to all purely agricultural communities, and
everywhere tends to beget in the husbandman a settled phlegmatic habit of
mind very different from the mobility, the alertness, the pliability of
character which the hazards and uncertainties of commerce and the sea
foster in the merchant and the sailor. The saturnine temperament of the
farmer is as naturally averse to change as the more mercurial spirit of the
trader and the seaman is predisposed to it. But the stereotyping of ideas
and of customs was carried further in Egypt than in most lands devoted to
husbandry by reason of the greater uniformity of the Egyptian seasons and
the more complete isolation of the country.
The general effect of these causes was to create a type of national character
which presented many points of resemblance to that of the Chinese. In both
we see the same inflexible strength of will, the same astonishing industry,
the same strange blend of humanity and savagery, the same obstinate
adherence to tradition, the same pride of race and of ancient civilization, the
same contempt for foreigners as for upstarts and barbarians, the same
patient outward submission to an alien rule combined with an unshakeable
inward devotion to native ideals. It was this conservative temper of the
people, bred in great measure of the physical nature of their land, which, so
to say, embalmed the memory of Osiris long after the corresponding figures
of Adonis and Attis had suffered decay. For while Egypt enjoyed profound
repose, the tides of war and conquest, of traffic and commerce, had for](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-30-320.jpg)
![centuries rolled over Western Asia, the native home of Adonis and Attis; and
if the shock of nationalities in this great meeting-ground of East and West
was favourable to the rise of new faiths and new moralities, it was in the
same measure unfavourable to the preservation of the old.
[pg 219]
Notes.](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-31-320.jpg)
![I. Moloch The King.
I cannot leave the evidence for the sacred character of Jewish kings544
without mentioning a suggestion which was made to me by my friend and
teacher the Rev. Professor R. H. Kennett. He thinks that Moloch, to whom
first-born children were burnt by their parents in the valley of Hinnom,
outside the walls of Jerusalem,545
may have been originally the human king
regarded as an incarnate deity. Certainly the name of Moloch, or rather
Molech (for so it is always written in the Massoretic text546
), is merely a
slightly disguised [pg 220] form of melech, the ordinary Hebrew word for
“king,” the scribes having apparently given the dreadful word the vowels of
bosheth, “shameful thing.”547
But it seems clear that in historical times the
Jews who offered these sacrifices identified Molech, not with the human
king, but with Jehovah, though the prophets protested against the custom
as an outrage on the divine majesty.548
If, however, these sacrifices were originally offered to or in behalf of the
human king, it is possible that they were intended to prolong his life and
strengthen his hands for the performance of those magical functions which
he was expected to discharge for the good of his people. The old kings of
Sweden answered with their heads for the fertility of the ground,549
and we
read that one of them, Aun or On by name, sacrificed nine of his sons to
Odin at Upsala in order that his own life might be spared. After the sacrifice
of his second son he received from the god an oracle that he should live as
long as he gave him one of his sons every tenth year. When he had thus
sacrificed seven sons, the ruthless father still lived, but was so feeble that
he could no longer walk and had to be carried in a chair. Then he offered up
his eighth son and lived ten years more, bedridden. After that he sacrificed
his ninth son, and lived ten years more, drinking out of a horn like a weaned
child. He now wished to sacrifice his last remaining son to Odin, but the
Swedes would not let him, so he died and was buried in a mound at
Upsala.550
In this Swedish tradition the king's children seem to have been
looked upon as substitutes offered to the god in place of their father, and
apparently this was also the current explanation of the slaughter of the first-
born in the later times of Israel.551
On that view the sacrifices were](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-32-320.jpg)
![vicarious, and therefore purely religious, being intended to propitiate a stern
and exacting deity. Similarly we read that when Amestris, wife of Xerxes,
was grown old, she sacrificed on her behalf twice seven noble children to
the [pg 221] earth god by burying them alive.552
If the story is true—and it
rests on the authority of Herodotus, a nearly contemporary witness—we
may surmise that the aged queen acted thus with an eye to the future
rather than to the past; she hoped that the grim god of the nether-world
would accept the young victims in her stead, and let her live for many years.
The same idea of vicarious suffering comes out in a tradition told of a
certain Hova king of Madagascar, who bore the sonorous name of
Andriamasinavalona. When he had grown sickly and feeble, the oracle was
consulted as to the best way of restoring him to health. “The following
result was the consequence of the directions of the oracle. A speech was
first delivered to the people, offering great honours and rewards to the
family of any individual who would freely offer himself to be sacrificed, in
order to the king's recovery. The people shuddered at the idea, and ran
away in different directions. One man, however, presented himself for the
purpose, and his offer was accepted. The sacrificer girded up his loins,
sharpened his knife, and bound the victim. After which, he was laid down
with his head towards the east, upon a mat spread for the purpose,
according to the custom with animals on such occasions, when the priest
appeared, to proceed with all solemnity in slaughtering the victim by cutting
his throat. A quantity of red liquid, however, which had been prepared from
a native dye, was spilled in the ceremony; and, to the amazement of those
who looked on, blood seemed to be flowing all around. The man, as might
be supposed, was unhurt; but the king rewarded him and his descendants
with the perpetual privilege of exemption from capital punishment for any
violation of the laws. The descendants of the man to this day form a
particular class, called Tay maty manota, which may be translated, ‘Not
dead, though transgressing.’ Instances frequently occur, of individuals of this
class appropriating bullocks, rice, and other things belonging to the
sovereign, as if they were their own, and escaping merely with a reprimand,
while a common person would have to suffer death, or be reduced to
slavery.”553
Sometimes, however, the practices intended to prolong the king's life seem
to rest on a theory of nutrition rather than of substitution; in other words,
the life of the victims, instead of being offered vicariously to a god, is
apparently supposed to pass directly into the body of the sacrificer, thus](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-33-320.jpg)
![refreshing his failing strength and prolonging his existence. So regarded, the
custom is magical rather than religious in character, since the desired effect
is thought to follow directly without the intervention of a deity. At all events,
it can be shown that sacrifices of this sort have been offered to prolong the
life of kings in other parts of the world. Thus in regard to [pg 222] some of
the negroes who inhabit the delta of the Niger we read that: “A custom
which formerly was practised by the Ibani, and is still prevalent among all
the interior tribes, consists in prolonging the life of a king or ancestral
representative by the daily, or possibly weekly, sacrifice of a chicken and
egg. Every morning, as soon as the patriarch has risen from his bed, the
sacrificial articles are procured either by his mother, head wife, or eldest
daughter, and given to the priest, who receives them on the open space in
front of the house. When this has been reported to the patriarch, he comes
outside and, sitting down, joins in the ceremony. Taking the chicken in his
hand, the priest first of all touches the patriarch's face with it, and
afterwards passes it over the whole of his body. He then cuts its throat and
allows the blood to drop on the ground. Mixing the blood and the earth into
a paste, he rubs it on the old man's forehead and breast, and this is not to
be washed off under any circumstances until the evening. The chicken and
the egg, also a piece of white cloth, are now tied on to a stick, which, if a
stream is in the near vicinity, is planted in the ground at the water-side.
During the carriage of these articles to the place in question, all the wives
and many members of the household accompany the priest, invoking the
deity as they go to prolong their father's life. This is done in the firm
conviction that through the sacrifice of each chicken his life will be
accordingly prolonged.”554
The ceremony thus described is, like so many other rites, a combination of
magic and religion; for whereas the prayers to the god are religious, the
passing of the victim over the king's body and the smearing of him with its
blood are magical, being plainly intended to convey to him directly, without
the mediation of any deity, the life of the fowl. In the following instances the
practices for prolonging the king's life seem to be purely magical. Among
the Zulus, at one of the annual feasts of first-fruits, a bull is killed by a
particular regiment. In slaughtering the beast they may not use spears or
sticks, but must break its neck or choke it with their bare hands. “It is then
burned, and the strength of the bull is supposed to enter into the king,
thereby prolonging his life.”555
Again, in an early Portuguese historian we
read of a Caffre king of East Africa that “it is related of this Monomotapa](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-34-320.jpg)
![that he has a house where he commands bodies of men who have died at
the hands of the law to be hung up, and where thus hanging all the
humidity [pg 223] of their bodies falls into vases placed underneath, and
when all has dropped from them and they shrink and dry up he commands
them to be taken down and buried, and with the fat and moisture in the
vases they say he makes ointments with which he anoints himself in order
to enjoy long life—which is his belief—and also to be proof against receiving
harm from sorcerers.”556
The Baganda of Central Africa used to kill men on various occasions for the
purpose of prolonging the king's life; in all cases it would seem to be
thought that the life of the murdered man was in some mysterious fashion
transferred to the king, so that the monarch received thereby a fresh
accession of vital energy. For example, whenever a particular royal drum
had a new skin put on it, not only was a cow killed to furnish the skin and
its blood run into the drum, but a man was beheaded and the spouting
blood from the severed neck was allowed to gush into the drum, “so that,
when the drum was beaten, it was supposed to add fresh life and vigour to
the king from the life of the slain man.”557
Again, at the coronation of a new
king, a royal chamberlain was chosen to take charge of the king's inner
court and to guard his wives. From the royal presence the chamberlain was
conducted, along with eight captives, to one of the human shambles; there
he was blindfolded while seven of the men were clubbed to death, only the
dull thud and crashing sound telling him of what was taking place. But when
the seven had been thus despatched, the bandages were removed from the
chamberlain's eyes and he witnessed the death of the eighth. As each man
was killed, his belly was ripped open and his bowels pulled out and hung
round the chamberlain's neck. These deaths were said to add to the King's
vigour and to make the chamberlain strong and faithful.558
Nor were these
the only human sacrifices offered at a king's coronation for the purpose of
strengthening the new monarch. When the king had reigned two or three
months, he was expected to hunt first a leopard and then a bushbuck. On
the night after the hunt of the bushbuck, one of the ministers of State
caught a man and brought him before the king in the dark; the king speared
him slightly, then the man was strangled and the body thrown into a
papyrus swamp, that it might never be found again. Another ceremony
performed about this time to confirm the king in his kingdom was to catch a
man, bind him, and bring him before the king, who wounded him slightly](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-35-320.jpg)
![with a spear. Then the man was put to death. These men were killed to
invigorate the king.559
[pg 224]
When a king of Uganda had reigned some time, apparently several years, a
ceremony was performed for the sake of prolonging his life. For this purpose
the king paid a visit—a fatal visit—to a chief of the Lung-fish clan, who bore
the title of Nankere and resided in the district of Busiro, where the tombs
and temples of the kings were situated. When the time for the ceremony
had been appointed, the chief chose one of his own sons, who was to die
that the king might live. If the chief had no son, a near relation was
compelled to serve as a substitute. The hapless youth was fed and clothed
and treated in all respects like a prince, and taken to live in a particular
house near the place where the king was to lodge for the ceremony. When
the destined victim had been feasted and guarded for a month, the king set
out on his progress from the capital. On the way he stopped at the temple
of the great god Mukasa; there he changed his garments, leaving behind
him in the temple those which he had been wearing. Also he left behind him
all his anklets, and did not put on any fresh ones, for he was shortly to
receive new anklets of a remarkable kind. When the king arrived at his
destination, the chief met him, and the two exchanged a gourd of beer. At
this interview the king's mother was present to see her son for the last time;
for from that moment the two were never allowed to look upon each other
again. The chief addressed the king's mother informing her of this final
separation; then turning to the king he said, “You are now of age; go and
live longer than your forefathers.” Then the chief's son was introduced. The
chief took him by the hand and presented him to the king, who passed him
on to the body-guard; they led him outside and killed him by beating him
with their clenched fists. The muscles from the back of the body of the
murdered youth were removed and made into two anklets for the king, and
a strip of skin cut from the corpse was made into a whip, which was kept in
the royal enclosure for special feasts. The dead body was thrown on waste
land and guarded against wild beasts, but not buried.560
When that ceremony was over, the king departed to go to another chief in
Busiro; but on the way thither he stopped at a place called Baka and sat
down under a great tree to play a game of spinning fruit-stones. It is a
children's game, but it was no child's play to the man who ran to fetch the
fruit-stones for the king to play with; for he was caught and speared to](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-36-320.jpg)
![death on the spot for the purpose of prolonging the king's life. After the
game had been played the king with his train passed on and lodged with a
certain princess till the anklets made from the muscles of the chief's
murdered son were ready for him to wear; [pg 225] it was the princess who
had to superintend the making of these royal ornaments.561
When all these ceremonies were over, the king made a great feast. At this
feast a priest went about carrying under his mantle the whip that had been
made from the skin of the murdered young man. As he passed through the
crowd of merrymakers, he would flick a man here and there with the whip,
and it was believed that the man on whom the lash lighted would be
childless and might die, unless he made an offering of either nine or ninety
cowrie shells to the priest who had struck him. Naturally he hastened to
procure the shells and take them to the striker, who, on receiving them,
struck the man on the shoulder with his hand, thus restoring to him the
generative powers of which the blow of the whip had deprived him. At the
end of the feast the drummers removed all the drums but one, which they
left as if they had forgotten it. Somebody in the crowd would notice the
apparent oversight and run after the drummers with the drum, saying, “You
have left one behind.” The thanks he received was that he was caught and
killed and the bones of his upper arm made into drumsticks for that
particular drum. The drum was never afterwards brought out during the
whole of the king's reign, but was kept covered up till the time came to
bring it out on the corresponding feast of his successor. Yet from time to
time the priest, who had flicked the revellers with the whip of human skin,
would dress himself up in a mantle of cow-hide from neck to foot, and
concealing the drumstick of human bones under his robe would go into the
king's presence, and suddenly whipping out the bones from his bosom
would brandish them in the king's face. Then he would as suddenly hide
them again, but only to repeat the manoeuvre. After that he retired and
restored the bones to their usual place. They were decorated with cowrie
shells and little bells, which jingled as he shook them at the king.562
The precise meaning of these latter ceremonies is obscure; but we may
suppose that just as the human blood poured into a drum was thought to
pass into the king's veins in the booming notes of the drum, so the clicking
of the human bones and the jingling of their bells were supposed to infuse
into the royal person the vigour of the murdered man. The purpose of
flicking commoners with the whip made of human skin is even more
obscure; but we may conjecture that the life or virility of every man struck](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-37-320.jpg)
![f
with the whip was supposed to be transmitted in some way to the king, who
thus recruited his vital, and especially his reproductive, energies at this
solemn feast. If I am right in my interpretation, all these Baganda [pg 226]
modes of strengthening the king and prolonging his life belonged to the
nutritive rather than to the vicarious type of sacrifice, from which it will
follow that they were magical rather than religious in character.
The same thing may perhaps be said of the wholesale massacres which
used to be perpetrated when a king of Uganda was ill. At these times the
priests informed the royal patient that persons marked by a certain physical
peculiarity, such as a cast of the eye, a particular gait, or a distinctive
colouring, must be put to death. Accordingly the king sent out his
catchpoles, who waylaid such persons in the roads and dragged them to the
royal enclosure, where they were kept until the tale of victims prescribed by
the priest was complete. Before they were led away to one of the eight
places of execution, which were regularly appointed for this purpose in
different parts of the kingdom, the victims had to drink medicated beer with
the king out of a special pot, in order that he might have power over their
ghosts, lest they should afterwards come back to torment him. They were
killed, sometimes by being speared to death, sometimes by being hacked to
pieces, sometimes by being burned alive. Contrary to the usual custom of
the Baganda, the bodies, or what remained of the bodies, of these
unfortunates were always left unburied on the place of execution.563
In what
way precisely the sick king was supposed to benefit by these massacres of
his subjects does not appear, but we may surmise that somehow the victims
were believed to give their lives for him or to him.
Thus it is possible that in Israel also the sacrifices of children to Moloch were
in like manner intended to prolong the life of the human king (melech)
either by serving as substitutes for him or by recruiting his failing energies
with their vigorous young life. But it is equally possible, and perhaps more
probable, that the sacrifice of the first-born children was only a particular
application of the ancient law which devoted to the deity the first-born of
every womb, whether of cattle or of human beings.564
[pg 227]](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-38-320.jpg)
![II. The Widowed Flamen.
§ 1. The Pollution of Death.
A different explanation of the rule which obliged the Flamen Dialis to resign
the priesthood on the death of his wife565
has been suggested by my friend
Dr. L. R. Farnell. He supposes that such a bereavement would render the
Flamen ceremonially impure, and therefore unfit to hold office.566
It is true
that the ceremonial pollution caused by death commonly disqualifies a man
for the discharge of sacred functions, but as a rule the disqualification is
only temporary and can be removed by seclusion and the observance of
purificatory rites, the length of the seclusion and the nature of the
purification varying with the degree of relationship in which the living stand
to the dead. Thus, for example, if one of the sacred eunuchs at Hierapolis-
Bambyce saw the dead body of a stranger, he was unclean for that day and
might not enter the sanctuary of the goddess; but next day after purifying
himself he was free to enter. But if the corpse happened to be that of a
relation he was unclean for thirty days and had to shave his head before he
might set foot within the holy precinct.567
Again, in the Greek island of Ceos
persons who had offered the annual sacrifices to their departed friends were
unclean for two days afterwards and might not enter a sanctuary; they had
to purify themselves with water.568
Similarly no one might go into the shrine
of Men Tyrannus for ten days after being in contact with the dead.569
Once
more, at Stratonicea in Caria a chorus of thirty noble boys, clad in white and
holding branches in their hands, used to sing a hymn daily in honour of Zeus
and Hecate; but if one of them were sick or had suffered a domestic
bereavement, he was for the time being excused, not permanently
excluded, from the [pg 228] performance of his sacred duties.570
On the
analogy of these and similar cases we should expect to find the widowed
Flamen temporarily debarred from the exercise of his office, not
permanently relieved of it.
However, in support of Dr. Farnell's view I would cite an Indian parallel
which was pointed out to me by Dr. W. H. R. Rivers. Among the Todas of the](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-40-320.jpg)
![Neilgherry Hills in Southern India the priestly dairyman (palol) is a sacred
personage, and his life, like that of the Flamen Dialis, is hedged in by many
taboos. Now when a death occurs in his clan, the dairyman may not attend
any of the funeral ceremonies unless he gives up office, but he may be re-
elected after the second funeral ceremonies have been completed. In the
interval his place must be taken by a man of another clan. Some eighteen or
nineteen years ago a man named Karkievan resigned the office of dairyman
when his wife died, but two years later he was re-elected and has held
office ever since. There have meantime been many deaths in his clan, but
he has not attended a funeral, and has not therefore had to resign his post
again. Apparently in old times a more stringent rule prevailed, and the
dairyman was obliged to vacate office whenever a death occurred in his
clan. For, according to tradition, the clan of Keadrol was divided into its two
existing divisions for the express purpose of ensuring that there might still
be men to undertake the office of dairyman when a death occurred in the
clan, the men of the one division taking office whenever there was a death
in the other.571
At first sight this case may seem exactly parallel to the case of the Flamen
Dialis and the Flaminica on Dr. Farnell's theory; for here there can be no
doubt whatever that it is the pollution of death which disqualifies the sacred
dairyman from holding office, since, if he only avoids that pollution by not
attending the funeral, he is allowed at the present day to retain his post. On
this analogy we might suppose that it was not so much the death of his wife
as the attendance at her funeral which compelled the Flamen Dialis to
resign, especially as we know that he was expressly forbidden to touch a
dead body or to enter the place where corpses were burned.572
But a closer inspection of the facts proves that the analogy breaks down at
some important points. For though the Flamen Dialis was forbidden to touch
a dead body or to enter a place where corpses were burned, he was
permitted to attend a funeral;573
so that there could hardly be any objection
to his attending the funeral of [pg 229] his wife. This permission clearly tells
against the view that it was the mere pollution of death which obliged him
to resign office when his wife died. Further, and this is a point of
fundamental difference between the two cases, whereas the Flamen Dialis
was bound to be married, and married too by a rite of special solemnity,574
there is no such obligation on the sacred dairyman of the Todas; indeed, if
he is married, he is bound to live apart from his wife during his term of
office.575
Surely the obligation laid on the Flamen Dialis to be married of](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-41-320.jpg)
![itself implies that with the death of his wife he necessarily ceased to hold
office: there is no need to search for another reason in the pollution of
death which, as I have just shown, does not seem to square with the
permission granted to the Flamen to attend a funeral. That this is indeed the
true explanation of the rule in question is strongly suggested by the further
and apparently parallel rule which forbade the Flamen to divorce his wife;
nothing but death might part them.576
Now the rule which enjoined that a
Flamen must be married, and the rule which forbade him to divorce his wife,
have obviously nothing to do with the pollution of death, yet they can hardly
be separated from the other rule that with the death of his wife he vacated
office. All three rules are explained in the most natural way on the
hypothesis which I have adopted, namely, that this married priest and
priestess had to perform in common certain rites which the husband could
not perform without his wife. The same obvious solution of the problem was
suggested long ago by Plutarch, who, after asking why the Flamen Dialis
had to lay down office on the death of his wife, says, amongst other things,
that “perhaps it is because she performs sacred rites along with him (for
many of the rites may not be performed without the presence of a married
woman), and to marry another wife immediately [pg 230] on the death of
the first would hardly be possible or decent.”577
This simple explanation of
the rule seems quite sufficient, and it would clearly hold good whether I am
right or wrong in further supposing that the human husband and wife in this
case represented a divine husband and wife, a god and goddess, to wit
Jupiter and Juno, or rather Dianus (Janus) and Diana;578
and that
supposition in its turn might still hold good even if I were wrong in further
conjecturing that of this divine pair the goddess (Juno or rather Diana) was
originally the more important partner.
However it is to be explained, the Roman rule which forbade the Flamen
Dialis to be a widower has its parallel among the Kotas, a tribe who, like the
Todas, inhabit the Neilgherry Hills of Southern India. For the higher Kota
priests are not allowed to be widowers; if a priest's wife dies while he is in
office, his appointment lapses. At the same time priests “should avoid
pollution, and may not attend a Toda or Badaga funeral, or approach the
seclusion hut set apart for Kota women.”579
Jewish priests were specially
permitted to contract the pollution of death for near relations, among whom
father, mother, son, daughter, and unmarried sister are particularly
enumerated; but they were forbidden to contract the pollution for strangers.](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-42-320.jpg)
![However, among the relations for whom a priest might thus defile himself a
wife is not mentioned.580
§ 2. The Marriage of the Roman Gods.
The theory that the Flamen Dialis and his wife personated a divine couple,
whether Jupiter and Juno or Dianus (Janus) and Diana, supposes a married
relation between the god and goddess, and so far it would certainly be
untenable if Dr. Farnell were right in assuming, on the authority of Mr. W.
Warde Fowler, that the Roman gods were celibate.581
On that subject,
however, Varro, the [pg 231] most learned of Roman antiquaries, was of a
contrary opinion. He not only spoke particularly of Juno as the wife of
Jupiter,582
but he also affirmed generally, in the most unambiguous
language, that the old Roman gods were married, and in saying so he
referred not to the religion of his own day, which had been modified by
Greek influence, but to the religion of the ancient Romans, his ancestors.583
Seneca ridiculed the marriage of the Roman gods, citing as examples the
marriages of Mars and Bellona, of Vulcan and Venus, of Neptune and
Salacia, and adding sarcastically that some of the goddesses were spinsters
or widows, such as Populonia, Fulgora, and Rumina, whose faded charms or
unamiable character had failed to attract a suitor.584
Again, the learned Servius, whose commentary on Virgil is a gold mine of
Roman religious lore, informs us that the pontiffs celebrated the marriage of
the infernal deity Orcus with very great solemnity;585
and for this statement
he would seem to have had the authority of the pontifical books themselves,
for he refers to them in the same connexion only a few lines before. As it is
in the highest degree unlikely that the pontiffs would solemnize any foreign
rites, we may safely assume that the marriage of Orcus was not borrowed
from Greek mythology, but was a genuine old Roman ceremony, and this is
all the more probable because Servius, our authority for the custom, has
recorded some curious and obviously ancient taboos which were observed
at the marriage and in the ritual of Ceres, the goddess who seems to have
been joined in wedlock to Orcus. One of these taboos forbade the use of
wine, the other forbade persons to name their father or daughter.586
[pg 232]](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-43-320.jpg)
![Further, the learned Roman antiquary Aulus Gellius quotes from “the books
of the priests of the Roman people” (the highest possible authority on the
subject) and from “many ancient speeches” a list of old Roman deities, in
which there seem to be at least five pairs of males and females.587
More
than that he proves conclusively by quotations from Plautus, the annalist
Cn. Gellius, and Licinius Imbrex that these old writers certainly regarded one
at least of the pairs (Mars and Nerio) as husband and wife;588
and we have
good ancient evidence for viewing in the same light three others of the
pairs. Thus the old annalist and antiquarian L. Cincius Alimentus, who
fought against Hannibal and was captured by him, affirmed in his work on
the Roman calendar that Maia was the wife of Vulcan;589
and as there was a
Flamen of Vulcan, who sacrificed to Maia on May Day,590
it is reasonable to
suppose that he was assisted in the ceremony by a Flaminica, his wife, just
as on my hypothesis the Flamen Dialis was assisted by his wife the
Flaminica. Another old Roman historian, L. Calpurnius Piso, who wrote in the
second century b.c., said that the name of Vulcan's wife was not Maia but
[pg 233] Majestas.591
In saying so he may have intended to correct what he
believed to be a mistake of his predecessor L. Cincius. Again, that Salacia
was the wife of Neptune is perhaps implied by Varro,592
and is positively
affirmed by Seneca, Augustine, and Servius.593
Again, Ennius appears to
have regarded Hora as the wife of Quirinus, for in the first book of his
Annals he declared his devotion to that divine pair.594
In fact, of the five
pairs of male and female deities cited by Aulus Gellius from the priestly
books and ancient speeches the only one as to which we have not
independent evidence that it consisted of a husband and wife is Saturn and
Lua; and in regard to Lua we know that she was spoken of as a mother,595
which renders it not improbable that she was also a wife. However,
according to some very respectable authorities the wife of Saturn was not
Lua, but Ops,596
so that we have two independent lines of proof that Saturn
was supposed to be married.
Lastly, the epithets “father” and “mother” which the Romans bestowed on
many of their deities597
are most naturally understood [pg 234] to imply
paternity and maternity; and if the implication is admitted, the inference
appears to be inevitable that these divine beings were supposed to exercise
sexual functions, whether in lawful marriage or in unlawful concubinage. As
to Jupiter in particular his paternity is positively attested by Latin
inscriptions, one of them very old, which describe Fortuna Primigenia, the
great goddess of Praeneste, as his daughter.598
Again, the rustic deity](https://image.slidesharecdn.com/10004849-250513080819-ea90fca5/85/A-Roadmap-For-Cognitive-Development-In-Humanoid-Robots-Vernon-44-320.jpg)
A Roadmap For Cognitive Development In Humanoid Robots Vernon
- 1. A Roadmap For Cognitive Development In Humanoid Robots Vernon download https://ebookbell.com/product/a-roadmap-for-cognitive- development-in-humanoid-robots-vernon-20009698 Explore and download more ebooks at ebookbell.com
- 3. Cognitive Systems Monographs Volume 11 Editors: Rüdiger Dillmann · Yoshihiko Nakamura · Stefan Schaal · David Vernon
- 4. David Vernon, Claes von Hofsten, and Luciano Fadiga A Roadmap for Cognitive Development in Humanoid Robots ABC
- 5. Rüdiger Dillmann, University of Karlsruhe, Faculty of Informatics, Institute of Anthropomatics, Humanoids and Intelligence Systems Laboratories, Kaiserstr. 12, 76131 Karlsruhe, Germany Yoshihiko Nakamura, Tokyo University Fac. Engineering, Dept. Mechano-Informatics, 7-3-1 Hongo, Bukyo-ku Tokyo, 113-8656, Japan Stefan Schaal, University of Southern California, Department Computer Science, Computational Learn- ing & Motor Control Lab., Los Angeles, CA 90089-2905, USA David Vernon, Department of Robotics, Brain and Cognitive Sciences, Italian Institute of Technology, Genoa, Italy Authors David Vernon Department of Robotics, Brain and Cognitive Sciences Italian Institute of Technology Genoa Italy E-mail: david@vernon.eu Claes von Hofsten Psykologisk institutt Universitetet i Oslo Oslo Norway E-mail: claes.von_hofsten@psyk.uu.se Luciano Fadiga Section of Human Physiology University of Ferrara Italy E-mail: fdl@unife.it and Department of Robotics, Brain and Cognitive Sciences Italian Institute of Technology Genoa ISBN 978-3-642-16903-8 e-ISBN 978-3-642-16904-5 DOI 10.1007/978-3-642-16904-5 Cognitive Systems Monographs ISSN 1867-4925 Library of Congress Control Number: 2010938643 c 2010 Springer-Verlag Berlin Heidelberg This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable for prosecution under the German Copyright Law. The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Typeset Cover Design: Scientific Publishing Services Pvt. Ltd., Chennai, India. Printed on acid-free paper 5 4 3 2 1 0 springer.com
- 6. Preface The work described in this book is founded on the premise that (a) cognition is the process by which an autonomous self-governing agent acts effectively in the world in which it is embedded, that (b) the dual purpose of cognition is to increase the agent’s repertoire of effective actions and its power to anticipate the need for and outcome of future actions, and that (c) development plays an essential role in the realization of these cognitive capabilities. Cognitive agents act in their world, typically with incomplete, uncertain, and time-varying sensory data. The chief purpose of cognition is to enable the selection of actions that are appropriate to the circumstances. However, the latencies inherent in the neural processing of sense data are often too great to allow effective action. Consequently, a cognitive agent must anticipate future events so that it can prepare the actions it may need to take. Furthermore, the world in which the agent is embedded is unconstrained so that it is not possible to predict all the circumstances it will experience and, hence, it is not possible to encapsulate a priori all the knowledge required to deal successfully with them. A cognitive agent then must not only be able to anticipate but it must also be able to learn and adapt, progressively increasing its space of possible actions as well as the time horizon of its prospective capabilities. In other words, a cognitive agent must develop. There are many implications of this stance. First, there must be some start- ing point for development — some phylogeny — both in terms of the initial capabilities and in terms of the mechanisms which support the developmental process. Second, there must be a developmental path — an ontogeny — which the agent follows in its attempts to develop an increased degree of prospection and a larger space of action. This involves several stages, from coordination of eye-gaze, head attitude, and hand placement when reaching, through to more complex exploratory use of action. This is typically achieved by dexter- ous manipulation of the environment to learn the affordances of objects in the context of one’s own developing capabilities. Third, since cognitive agents rarely operate in isolation and since the world with which they interact typi- cally includes other cognitive agents, there is the question of how a cognitive
- 7. VI Preface agent can share with other agents the knowledge it has learned. Since what an agent knows is based on its history of experiences in the world, the mean- ing of any shared knowledge depends on a common mode of experiencing the world. In turn, this implies that the shared knowledge is predicated upon the agents having a common morphology and, in the case of human-robot interaction, a common humanoid form. The roadmap set out in this book targets specifically the development of cognitive capabilities in humanoid robots. It identifies the necessary and hopefully sufficient conditions that must exist to allow this development. It has been created by bringing together insights from four areas: enactive cognitive science, developmental psychology, neurophysiology, and cognitive modelling. Thus, the roadmap builds upon what is known about the develop- ment of natural cognitive systems and what is known about computational modelling of artificial cognitive agents. In doing so, it identifies the essential principles of a system that can develop cognitive capabilities and it shows how these principles have been applied to the state-of-the-art humanoid robot: the iCub. The book is organized as follows. Chapter 1 presents a conceptual frame- work that forms the foundation of the book. It identifies the broad stance taken on cognitive systems — emergent embodied systems that develop cogni- tive skills as a result of their action in the world — and it draws out explicitly the consequences of adopting this stance. Chapter 2 begins by discussing the importance of action as the organizing principle in cognitive behaviour, a theme that will recur repeatedly throughout the book. It then addresses the phylogeny of human infants and, in particular, it considers the innate capa- bilities of pre-natal infants and how these develop before and just after birth. Chapter 3 then details how these capabilities develop in the first couple of years of life, focussing on the interdependence of perception and action. In do- ing so, it develops the second recurrent theme of the book: the central role of prospective capabilities in cognition. Chapter 4 explores the neurophysiology of perception and action, delving more deeply into the way that the interde- pendency of perception and action is manifested in the primate brain. While Chapters 2 – 4 provide the biological inspiration for the design of an entity that can develop cognitive capabilities, Chapter 5 surveys recent attempts at building artificial cognitive systems, focussing on cognitive architectures as the basis for development. Chapter 6 then presents a complete roadmap that uses the phylogeny and ontogeny of natural systems as well as insights gained from computational models and cognitive architectures to define the innate capabilities with which the humanoid robot must be equipped so that it is capable of ontogenetic development. The roadmap includes a series of sce- narios that can be used to drive the robot’s developmental progress. Chapter 7 provides an overview of the iCub humanoid robot and it describes the use of the the roadmap in the realization of the iCub’s own cognitive architec- ture. Chapter 8 concludes by setting out an agenda for future research and
- 8. Preface VII addressing the most pressing issues that will advance our understanding of cognitive systems, artificial and natural. Dublin, Uppsala, and Ferrara David Vernon August 2010 Claes von Hofsten Luciano Fadiga
- 9. Acknowledgements This book is based on the results of the RobotCub research project, the goal of which was to develop the iCub: an open and widely-adopted humanoid robot for cognitive systems research. This project was supported by the Eu- ropean Commission, Project IST-004370, under Strategic Objective 2.3.2.4: Cognitive Systems and we gratefully acknowledge the funding that made this research possible. In particular, we would like to thank Hans-Georg Stork, European Com- mission, for his support, insight, and unfailing belief in the project. We would also like to acknowledge the contributions made by the five reviewers of the project over its five-and-a-half year lifetime: Andreas Birk, Joanna Bryson, Bob Damper, Peter Ford Dominey, and Raul Rojas. Their collective suggestions helped greatly in navigating uncharted waters. We are indebted to the members of the project’s International Advisory Board — Rodney Brooks, MIT, Gordon Cheng, Technical University of Mu- nich, Jürgen Konczak, University of Minnesota, Hideki Kozima, CRL, and Yasuo Kuniyoshi, University of Tokyo — for providing valuable advice and moral support. Over one hundred people were involved in the creation of the iCub and it is impossible to acknowledge the contribution each made to the work that is described in this book. However, certain key individuals in each of the eleven institutes that participated in the research played a pivotal role is bringing the five year project to a successful conclusion. It is a pleasure to acknowledge their contributions. Giulio Sandini, Italian Institute of Technology and University of Genoa, was the mastermind behind the project and he was the first to see the need for a common humanoid robot platform to support research in embodied cognitive systems and the benefits of adopting an open-systems policy for software and hardware development. Giorgio Metta, Italian Institute of Technology and University of Genoa, inspired many of the design choices and provided leadership, guidance, and a level of commitment to the project that was crucial for its success.
- 10. X Acknowledgements Paul Fitzpatrick, Lorenzo Natale and Francesco Nori, Italian Institute of Technology and University of Genoa, together formed an indispensible team whose wide-ranging contributions to the software and hardware formed the critical core of the iCub. José Santos-Victor and Alex Bernardino, IST Lisbon, set the benchmark early on with their design of the iCub head and their work on learning affor- dances, computational attention, gaze control, and hand-eye coordination. Francesco Becchi, Telerobot S.r.l., provided the industry-strength know- how which ensured that mechanical components of the iCub were designed, fabricated, assembled, and tested to professional standards. Rolf Pfeifer, University of Zurich, provided inspiration for the tight re- lationship between a system’s embodiment and its cognitive behaviour, a relationship that manifests itself in several ways in the the design of the iCub. Kerstin Rosander, University of Uppsala, was instrumental in keeping the work on developmental psychology in focus and on track. Laila Craighero, University of Ferrara, provided many of the key insights from neurophysiology which guided our models of cognitive development. Paolo Dario and Cecilia Laschi, Scuola S. Anna, Pisa contributed expertise in robot control systems for visual servo control and tracking. Kerstin Dautenhahn, Chrystopher Nehaniv, University of Hertfordshire, provided guidance on social human-robot interaction. Darwin Caldwell and Nikos Tsagarakis, Italian Institute of Technology and University of Sheffield, were responsible of the success of the design of the iCub’s torso and legs. Aude Billard and Auke Ijspeert, Ecole Polytechnique Federal de Lausanne, developed the software for imitation-based learning and gait control, respec- tively. To the countless other people we haven’t named — Commission officers, technicians, professors, Ph.D. and M.Sc. students, research assistants, office adminstrators, post-docs — a heart-felt thank you. This book reflects only a small part of the RobotCub project but it wouldn’t have been possible without the collective effort of everyone involved in creating the iCub. Finally, a big thank you to Keelin for her painstaking work proofreading the manuscript.
- 11. Contents 1 A Conceptual Framework for Developmental Cognitive Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Cognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 Enaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.3.1 Enaction and Development . . . . . . . . . . . . . . . . . . . . . . . 5 1.3.2 Enaction and Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . 7 1.3.3 Phylogeny and Ontogeny: The Complementarity of Structural Determination and Development . . . . . . . . . 8 1.4 Embodiment: The Requirements and Consequences of Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.5 Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2 Pre-natal Development and Core Abilities . . . . . . . . . . . . . . . 13 2.1 Action as the Organizing Principle in Cognitive Behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.2 Prenatal Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 2.2.1 Morphological Pre-structuring . . . . . . . . . . . . . . . . . . . . 16 2.2.2 Pre-structuring of the Motor System . . . . . . . . . . . . . . . 16 2.2.3 Pre-structuring of the Perceptual System . . . . . . . . . . . 18 2.2.4 Forming Functional Systems . . . . . . . . . . . . . . . . . . . . . . 20 2.3 Core Abilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.3.1 Core Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.3.2 Core Motives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 2.4.1 Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 2.4.2 Prenatal Development . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 2.4.3 Core Abilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
- 12. XII Contents 3 The Development of Cognitive Capabilities in Infants . . . . 29 3.1 The Development of Perception . . . . . . . . . . . . . . . . . . . . . . . . . 29 3.2 Visual Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 3.2.1 Space Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 3.2.2 Object Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 3.3 Acquiring Predictive Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 3.3.1 Development of Posture and Locomotion . . . . . . . . . . . 35 3.3.2 Development of Looking . . . . . . . . . . . . . . . . . . . . . . . . . . 40 3.3.3 Development of Reaching and Manipulation . . . . . . . . 44 3.3.4 Development of Social Abilities . . . . . . . . . . . . . . . . . . . 54 3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 3.4.1 The Basis for Development . . . . . . . . . . . . . . . . . . . . . . . 57 3.4.2 Visual Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 3.4.3 Posture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 3.4.4 Gaze. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 3.4.5 Reaching and Grasping . . . . . . . . . . . . . . . . . . . . . . . . . . 61 3.4.6 Manipulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 3.4.7 Social Abilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 4 What Neurophysiology Teaches Us About Perception and Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 4.1 The Premotor Cortex of Primates Encodes Actions and Not Movements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 4.2 The System for Grasping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 4.3 The Distributed Perception of Space . . . . . . . . . . . . . . . . . . . . . 70 4.4 Perception Depends on Action . . . . . . . . . . . . . . . . . . . . . . . . . . 72 4.5 Action and Selective Attention . . . . . . . . . . . . . . . . . . . . . . . . . . 73 4.6 Structured Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 4.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 4.7.1 Grasping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 4.7.2 Spatial Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 4.7.3 Perception-Action Dependency . . . . . . . . . . . . . . . . . . . . 78 4.7.4 Structured Interactions. . . . . . . . . . . . . . . . . . . . . . . . . . . 79 4.7.5 Selective Attention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 5 Computational Models of Cognition . . . . . . . . . . . . . . . . . . . . . 81 5.1 The Three Paradigms of Cognition . . . . . . . . . . . . . . . . . . . . . . 81 5.1.1 The Cognitivist Paradigm . . . . . . . . . . . . . . . . . . . . . . . . 85 5.1.2 The Emergent Paradigm . . . . . . . . . . . . . . . . . . . . . . . . . 86 5.1.3 The Hybrid Paradigm. . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 5.1.4 Relative Strengths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 5.2 Cognitive Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 5.2.1 The Cognitivist Perspective on Cognitive Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
- 13. Contents XIII 5.2.2 The Emergent Perspective on Cognitive Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 5.2.3 Desirable Characteristics of a Cognitive Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 5.2.4 A Survey of Cognitive Architectures . . . . . . . . . . . . . . . 95 5.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 6 A Research Roadmap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 6.1 Phylogeny . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 6.1.1 Guidelines from Enaction. . . . . . . . . . . . . . . . . . . . . . . . . 103 6.1.2 Guidelines from Developmental Psychology . . . . . . . . . 103 6.1.3 Guidelines from Neurophysiology . . . . . . . . . . . . . . . . . . 105 6.1.4 Guidelines from Computational Modelling . . . . . . . . . . 105 6.1.5 A Summary of the Phylogenetic Guidelines for the Development of Cognition in Artificial Systems. . . . . . 108 6.2 Ontogeny . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 6.2.1 Guidelines from Developmental Psychology . . . . . . . . . 110 6.2.2 Scenarios for Development . . . . . . . . . . . . . . . . . . . . . . . . 111 6.2.3 Scripted Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 7 The iCub Cognitive Architecture . . . . . . . . . . . . . . . . . . . . . . . . 121 7.1 The iCub Humanoid Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 7.1.1 The iCub Mechatronics . . . . . . . . . . . . . . . . . . . . . . . . . . 122 7.1.2 The iCub Middleware . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 7.2 The iCub Cognitive Architecture . . . . . . . . . . . . . . . . . . . . . . . . 125 7.2.1 Embodiment: The iCub Interface . . . . . . . . . . . . . . . . . . 127 7.2.2 Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 7.2.3 Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 7.2.4 Anticipation Adaptation . . . . . . . . . . . . . . . . . . . . . . . 138 7.2.5 Motivation: Affective State . . . . . . . . . . . . . . . . . . . . . . . 141 7.2.6 Autonomy: Action Selection . . . . . . . . . . . . . . . . . . . . . . 142 7.2.7 Software Implementation . . . . . . . . . . . . . . . . . . . . . . . . . 142 7.3 The iCub Cognitive Architecture vs. the Roadmap Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 7.3.1 Embodiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 7.3.2 Perception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 7.3.3 Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 7.3.4 Anticipation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 7.3.5 Adaptation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 7.3.6 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 7.3.7 Autonomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 7.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
- 14. XIV Contents 8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 8.1 Multiple Mechanisms for Anticipation . . . . . . . . . . . . . . . . . . . . 155 8.2 Prediction, Reconstruction, and Action: Learning Affordances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 8.3 Object Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 8.4 Multi-modal and Hierarchical Episodic Memory . . . . . . . . . . . 157 8.5 Generalization and Model Generation . . . . . . . . . . . . . . . . . . . . 157 8.6 Homeostasis and Development . . . . . . . . . . . . . . . . . . . . . . . . . . 158 A Catalogue of Cognitive Architectures . . . . . . . . . . . . . . . . . . . . 159 A.1 Cognitivist Cognitive Architectures . . . . . . . . . . . . . . . . . . . . . . 160 A.1.1 The Soar Cognitive Architecture . . . . . . . . . . . . . . . . . . 160 A.1.2 EPIC — Executive Process Interactive Control . . . . . 161 A.1.3 ACT-R — Adaptive Control of Thought - Rational . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 A.1.4 The ICARUS Cognitive Architecture. . . . . . . . . . . . . . . 165 A.1.5 ADAPT — A Cognitive Architecture for Robotics . . . 167 A.1.6 The GLAIR Cognitive Architecture . . . . . . . . . . . . . . . . 168 A.1.7 CoSy Architecture Schema . . . . . . . . . . . . . . . . . . . . . . . 170 A.2 Emergent Cognitive Architectures . . . . . . . . . . . . . . . . . . . . . . . 174 A.2.1 Autonomous Agent Robotics . . . . . . . . . . . . . . . . . . . . . . 174 A.2.2 A Global Workspace Cognitive Architecture . . . . . . . . 175 A.2.3 Self-directed Anticipative Learning . . . . . . . . . . . . . . . . 178 A.2.4 A Self-Affecting Self-Aware (SASE) Cognitive Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 A.2.5 Darwin: Neuromimetic Robotic Brain-Based Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 A.2.6 The Cognitive-Affective Architecture. . . . . . . . . . . . . . . 183 A.3 Hybrid Cognitive Architectures. . . . . . . . . . . . . . . . . . . . . . . . . . 186 A.3.1 A Humanoid Robot Cognitive Architecture . . . . . . . . . 186 A.3.2 The Cerebus Architecture . . . . . . . . . . . . . . . . . . . . . . . . 187 A.3.3 Cog: Theory of Mind . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 A.3.4 Kismet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190 A.3.5 The LIDA Cognitive Architecture . . . . . . . . . . . . . . . . . 192 A.3.6 The CLARION Cognitive Architecture . . . . . . . . . . . . . 194 A.3.7 The PACO-PLUS Cognitive Architecture . . . . . . . . . . . 196 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
- 15. Chapter 1 A Conceptual Framework for Developmental Cognitive Systems 1.1 Introduction This book addresses the central role played by development in cognition. We are in- terested in particular in applying our knowledge of development in natural cognitive systems, i.e. human infants, to the problem of creating artificial cognitive systems in the guise of humanoid robots. Thus, our subject matter is cognition, development, and humanoid robotics. These three threads are woven together to form a roadmap that when followed will enable the instantiation and development of an artificial cognitive system. However, to begin with, we must be clear what we mean by the term cognition so that, in turn, we can explain the pivotal role of development and the central relevance of humanoid embodiment. In the following, we present a conceptual framework that identifies and explains the broad stance we take on cognitive systems — emergent embodied systems that develop cognitive skills as a result of their action in the world — and that draws out explicitly the theoretical and practical consequences of adopting this stance.1 We begin by considering the operational characteristics of a cognitive system, focussing on the purpose of cognition rather than debating the relative merits of competing paradigms of cognition. Of course, such a debate is important because it allows us to understand the pre-conditions for cognition so, once we have estab- lished the role cognition plays and see why it is important, we move on to elaborate on these pre-conditions. In particular, we introduce the underlying framework of enaction which we adopt as the basis for the research described in this book. By working through the implications of the enactive approach to cognition, the central role of development in cognition becomes clear, as do several other key issues such as the crucial role played by action, the inter-dependence of percep- tion and action, and the consequent constructivist nature of the cognitive system’s knowledge. 1 This chapter is based directly on a study of enaction as a framework for development in cognitive robotics [385]. The original paper contains additional technical details relating to enactive systems which are not strictly required here. Readers who are interested in delving more deeply into enaction are encouraged to refer to the original. D. Vernon et al.: A Roadmap for Cogn. Develop. in Humanoid Robots, COSMOS 11, pp. 1–11. springerlink.com c Springer-Verlag Berlin Heidelberg 2010
- 16. 2 1 A Conceptual Framework for Developmental Cognitive Systems The framework of enaction provides the foundation for subsequent chapters which deal with the phylogeny and the ontogeny of natural cognitive systems — their initial capabilities and subsequent development — and the application of what we learn from these to the realization of an artificial cognitive system in the form of a humanoid robot. 1.2 Cognition Cognitive systems anticipate, assimilate, and adapt. In doing so, they learn and de- velop [387]. Cognitive systems anticipate future events when selecting actions, they subsequently learn from what actually happens when they do act, and thereby they modify subsequent expectations and, in the process, they change how the world is perceived and what actions are possible. Cognitive systems do all of this au- tonomously. The adaptive, anticipatory, autonomous viewpoint reflects the position of Freeman and Núñez who, in their book Reclaiming Cognition [105], assert the primacy of action, intention, and emotion in cognition. In the past, however, cog- nition was viewed in a very different light as a symbol-processing module of mind concerned with rational planning and reasoning. Today, however, this is changing and even proponents of these early approaches now see a much tighter relationship between perception, action, and cognition (e.g. see [7, 214]). So, if cognitive systems anticipate, assimilate, and adapt, if they develop and learn, the first question to ask is why do they do this? The subsequent question is how do they do it? The remainder of this section is devoted to the first question and the rest of the book addresses the latter. The view of cognition taken in this book is that cognition is the process whereby an autonomous self-governing system acts effectively in the world in which it is embedded [237]. However, in natural systems, the latencies inherent in the neu- ral processing of sense data are too great to allow effective action. This is one of the primary reasons a cognitive agent must anticipate future events: so that it can prepare the actions it may need to take. In addition, there are also limitations im- posed by the environment and the cognitive system’s body. To perform an action, one needs to have the relevant body part in a certain place at a certain time. In a dynamic environment that is constantly changing and with a body that takes time to move, this requires preparation and prediction. Furthermore, the world in which the agent is embedded is unconstrained and the sensory data which is available to the cognitive system is not only ‘out-of-date’ but it is also uncertain and incomplete. Consequently, it is not possible to encapsulate a priori all the knowledge required to deal successfully with the circumstances it will experience so that it must also be able to adapt, progressively increasing its space of possible actions as well as the time horizon of its prospective capabilities. It must do this, not as a reaction to external stimuli but as a self-generated process of proactive understanding. This process is what we mean by development. In summary, and as noted in the Preface, the position being set out in this book is that (a) cognition is the process by which an autonomous self-governing agent acts effectively in the world in which it is em- bedded, that (b) the dual purpose of cognition is to increase the agent’s repertoire
- 17. 1.3 Enaction 3 of effective actions and its power to anticipate the need for and outcome of future actions, and that (c) development plays an essential role in the realization of these cognitive capabilities. We will now introduce a framework which encapsulates all these considerations. 1.3 Enaction There are many alternative perspectives on cognition: what it is, why it is necessary, and how it is achieved. We have already argued that cognition arises from an agent’s need to compensate for latencies in neural processing by anticipating what may be about to happen and by preparing its actions accordingly. So we can agree fairly easily what cognition is — a process of anticipating events and acting appropriately and effectively — and why it is necessary — to overcome the physical limitations of biological brains and the limitations of bodily movements operating in a dynamic environment. The difficulty arises when we consider how cognition is achieved. There are several competing theories of cognition, each of which makes its own set of assumptions. Here, we wish to focus on one particularly important paradigm — enaction — and pick out its most salient aspects in order to provide a sound theoretical foundation for the role of development in cognition [235, 236, 237, 238, 359, 380, 382, 381, 400] . The principal idea of enaction is that a cognitive system develops it own under- standing of the world around it through its interactions with the environment. Thus, enaction entails that the cognitive system operates autonomously and that it gener- ates its own models of how the world works. When dealing with enactive systems, there are five key elements to consider [280, 373]: 1. Autonomy 2. Embodiment 3. Emergence 4. Experience 5. Sense-making Autonomy is the self-maintaining organizational characteristic of living creatures that enables them to use their own capacities to manage their interactions with the world, and with themselves, in order to remain viable [61, 108]. This simply means that the system is entirely self-governing and self-regulating: it is not controlled by any outside agency and this allows it to stand apart from the rest of the environment and operate independently of it. That’s not to say that the system isn’t influenced by the world around it, but rather that these influences are brought about through interactions that don’t threaten the autonomous operation of the system. The second element of enaction is the idea of embodiment. For our purposes here, embodiment means that the system must exist in the world as a physical entity which can interact directly with the environment. This means the system can act on things in the world around it and they, in turn, can act on the system. These things can be inanimate objects or animate agents, cognitive or not. As it happens, there
- 18. 4 1 A Conceptual Framework for Developmental Cognitive Systems are some subtle distinctions which can be made about different types of embodiment and we will return to this topic later in the chapter in Section 1.4. The element of emergence refers to the manner in which cognition arises in the system. Specifically, it refers to the laws and mechanisms which govern the behaviour of the component parts of the system. In an emergent system, the behaviour we call cognition arises from the dynamic interplay between the com- ponents and the laws and mechanisms we mentioned govern only the behaviour of the component parts; they don’t specify the behaviour of the interplay between the components. Thus, behaviour emerges indirectly because of the internal dynamics. Crucially, these internal dynamics must maintain the autonomy of the system and, as we will see shortly, they also condition the experiences of the system through their embodiment in a specific structure. Experience is the fourth element of enaction. This is nothing more than the cog- nitive system’s history of interaction with the world around it: the actions it takes in the environment and the actions arising in the environment which impinge on the cognitive system. These interactions don’t control the system (otherwise it wouldn’t be autonomous) but they do trigger changes in the state of the system. The changes that can be triggered are structurally determined: they depend on the system struc- ture, i.e. the embodiment of the self-organizational principles that make the system autonomous. This structure is also referred to as the system’s phylogeny: the innate capabilities of an autonomous system with which it is equipped at the outset and which form the basis for its continued existence. The experience of the systems — its history of interactions — involving structural coupling between the system and its environment is referred to as its ontogeny. Finally, we come to the fifth and, arguably, the most important element of en- action: sense-making. This term refers to the relationship between the knowledge encapsulated by a cognitive system and the interactions which gave rise to it. In particular, it refers to the idea that this emergent knowledge is generated by the sys- tem itself and that it captures some regularity or lawfulness in the interactions of the system, i.e. its experience. However, the sense it makes is dependent on the way in which it can interact: its own actions and its perceptions of the environment’s action on it. Since these perceptions and actions are the result of an emergent dy- namic process that is first and foremost concerned with maintaining the autonomy and operational identity of the system, these perceptions and actions are unique to the system itself and the resultant knowledge makes sense only insofar as it con- tributes to the maintenance of the system’s autonomy. This ties in neatly with our view of cognition, the role of which is to anticipate events and increase the space of actions in which a system can engage. By making sense of its experience, the cogni- tive system is constructing a model that has some predictive value, exactly because it captures some regularity or lawfulness in its interactions. This self-generated model of the system’s experience lends the system greater flexibility in how it in- teracts in the future. In other words, it endows the system with a larger repertoire of possible actions that allow richer interactions, increased perceptual capacity, and the possibility of constructing even better models that encapsulate knowlege with even greater predictive power. And so it goes, in a virtuous circle. Note that this
- 19. 1.3 Enaction 5 Fig. 1.1 Maturana and Varela’s ideograms to denote structurally-determined autopoietic and organizationally-closed systems. The arrow circle denotes the autonomy, self-organization, and self-production of the system, the rippled line the environment, and the bi-directional half-lines the mutual perturbation — structural coupling — between the two. sense-making and the resultant knowledge says nothing at all about what is really out there in the environment; it doesn’t have to. All it has to do is make sense for the continued existence and autonomy of the cognitive system. Sense-making is actu- ally the source of the term enaction. In making sense of its experience, the cognitive system is somehow bringing out through its actions — enacting — what is important for the continued existence of the system. This enaction is effected by the system as it is embedded in its environment, but as an autonomous entity distinct from the environment, through an emergent process of making sense of its experience. This sense-making is, in fact, cognition [108]. The founders of the enactive approach, Maturana and Varela, introduced a dia- grammatic way of conveying the self-organizing and self-maintaining autonomous nature of an enactive system, perturbing and being perturbed by its environment [237]: see Fig. 1.1. The arrowed circle denotes the autonomy and self-organization of the system, the rippled line the environment, and the bi-directional half-arrows the mutual perturbation. 1.3.1 Enaction and Development So what has all this to do with development? Our position in this book is that learn- ing arises as a consequence of the interaction between the cognitive agent and the world around it, whereas development arises from learning through a process of in- teraction of the cognitive agent with itself. We remarked above that the process of sense-making forms a virtuous circle in that the self-generated model of the sys- tem’s experience provides a larger repertoire of possible actions, richer interactions, increased perceptual capacity, and potentially better self-generated models, and so on. Recall also our earlier remarks that the cognitive system’s knowledge is rep- resented by the state of the system. When this state is embodied in the system’s central nervous system, the system has much greater plasticity in two senses: (a) the nervous system can accommodate a much larger space of possible associations be- tween system-environment interactions, and (b) it can accommodate a much larger space of potential actions. Consequently, the process of cognition involves the sys- tem modifying its own state, specifically its central nervous system, as it enhances
- 20. 6 1 A Conceptual Framework for Developmental Cognitive Systems its predictive capacity and its action capabilities. This is exactly what we mean by development. This generative (i.e. self-constructed) autonomous learning and devel- opment is one of the hallmarks of the enactive approach [280, 108]. Development, then, is identically the cognitive process of establishing and en- larging the possible space of mutually-consistent couplings in which a system can engage (or, perhaps more appropriately, which it can withstand without compro- mising its autonomy). The space of perceptual possibilities is predicated not on an absolute objective environment, but on the space of possible actions that the system can engage in whilst still maintaining the consistency of the coupling with the en- vironment. These environmental perturbations don’t control the system since they are not components of the system (and, by definition, don’t play a part in the self- organization) but they do play a part in the ontogenetic development of the system. Through this ontogenetic development, the cognitive system develops its own epis- temology, i.e. its own system-specific history- and context-dependent knowledge of its world, knowledge that has meaning exactly because it captures the consis- tency and invariance that emerges from the dynamic self-organization in the face of environmental coupling. Again, it comes down to the preservation of auton- omy, but this time doing so in an every increasing space of autonomy-preserving couplings. This process of development is achieved through self-modification by virtue of the presence of a central nervous system: not only does environment perturb the system (and vice versa) but the system also perturbs itself and the central nervous system adapts as a result. Consequently, the system can develop to accommodate a much larger space of effective system action. This is captured in a second ideogram of Maturana and Varela (see Fig. 1.2) which adds a second arrow circle to the ideogram to depict the process of development through self-perturbation and self- modification. In essence, development is autonomous self-modification and requires the existence of a viable phylogeny, including a nervous system, and a suitable ontogeny. Fig. 1.2 Maturana and Varela’s ideograms to denote structurally-determined autopoietic and operationally-closed systems. The diagram (denotes an organizationally-closed autonomous system with a central nervous system. This system is capable of development by means of self-modification of its nervous system, so that it can accommodate a much larger space of effective system action.
- 21. 1.3 Enaction 7 1.3.2 Enaction and Knowledge Let us now move on to discuss in a little more detail the nature of the knowledge that an enactive cognitive system constructs. This knowledge is built on sensorimotor as- sociations, achieved initially by exploration of what the world offers. However, this is only the beginning. The enactive system uses the knowledge gained to form new knowledge which is then subjected to empirical validation to see whether or not it is warranted (we, as enactive beings, imagine many things but not everything we imag- ine is plausible or corresponds well with reality). One of the key issues in cognition is the importance of internal simulation in accelerating the scaffolding of this early developmentally-acquired sensorimotor knowledge to provide a means to predict future events, to reconstruct or explain observed events (constructing a causal chain leading to that event), or to imagine new events [132, 144, 337]. Naturally, there is a need to focus on (re-)grounding predicted, explained, or imagined events in expe- rience so that the system can do something new and interact with the environment in a new way. If the cognitive system wishes or needs to share this knowledge with other cognitive systems or communicate with other cognitive systems, it will only be possible if they have shared a common history of experiences and if they have a similar phylogeny and a compatible ontogeny. In other words, the meaning of the knowledge that is shared is negotiated and agreed by consensus through interaction. When there are two or more cognitive agents involved, interaction is a shared activity in which the actions of each agent influence the actions of the other agents engaged in the same interaction, resulting in a mutually constructed pattern of shared behaviour [275]. Again, Maturana and Varela introduce a succinct diagram- matic way of of conveying this coupling between cognitive agent and the develop- ment it engenders [238]: see Fig. 1.3. Such mutually-constructedpatterns of comple- mentary behaviour is also emphasized in Clark’s notion of joint action [64]. Thus, explicit meaning is not necessary for anything to be communicated in an interac- tion, it is simply important that the agents are mutually engaged in a sequence of actions. Meaning emerges through shared consensual experience mediated by in- teration. The research programme encapsulated in this roadmap is based on this foundational principle of interaction. The developmental progress of imitation fol- lows tightly that of the development of other interactive and communicative skills, Fig. 1.3 Maturana and Varela’s ideogram to denote the development engendered by interac- tion between cognitive systems
- 22. Other documents randomly have different content
- 23. heat of the sun, and with a shade of banana leaves for her eyes. There, dripping with sweat in the burning sun and coated with mud to the hips and over the elbows, she toils to set the younger women a good example. Moreover, as in every other occupation, the kaliths, the gods, must also be invoked, and who could be better fitted for the discharge of so important a duty than the Mother of the House?”522 It seems clear that in any agricultural people who, like the Pelew Islanders, retain mother-kin and depute the labours of husbandry to women, the conception of a great Mother Goddess, the divine source of all fertility, might easily originate. Perhaps the same social and industrial conditions may have combined to develop the great Mother Goddesses of Western Asia and Egypt. But in the Pelew Islands women have yet another road to power. For some of them are reputed to be the wives of gods, and act as their oracular mouthpieces. Such prophetesses are called Amlaheys, and no surprise is felt when one of them is brought to bed. Her child passes for the offspring of the god, her divine husband, and goes about with his hair hanging loose in token of his superhuman parentage. It is thought that no mortal man would dare to intrigue with one of these human wives of a god, since the jealous deity would surely visit the rash culprit with deadly sickness and a lingering decline.523 But in these islands men as well as women are often possessed by a deity and speak in his name. Under his inspiration they mimic, often with great histrionic skill, the particular appearance and manner which are believed to be characteristic of the indwelling divinity. These inspired men (Korongs) usually enjoy great consideration and exert a powerful influence over the whole community. They always acquire wealth in the exercise of their profession. When they are not themselves chiefs, they are treated as chiefs or even preferred to them. In not a few places the deity whom [pg 208] they personate is also the political head of the land; and in that case his inspired priest, however humble his origin, ranks as a spiritual king and rules over all the chiefs. Indeed we are told that, with the physical and intellectual decay of the race, the power of the priests is more and more in the ascendant and threatens, if unchecked, to develop before long into an absolute theocracy which will swallow up every other form of government.524 Thus the present, or at least the recent, state of society and religion in the Pelew Islands presents some interesting parallels to the social and religious condition of Western Asia and Egypt in early days, if the conclusions reached in this work are correct. In both regions we see a society based on mother-kin developing a religion in which goddesses of the clan originally
- 24. f f f occupied the foremost place, though in later times, as the clans coalesced into states, the old goddesses have been rivalled and to some extent supplanted by the new male gods of the enlarged pantheon. But in the religion of the Pelew Islanders, as in that of the Khasis and the ancient Egyptians, the balance of power has never wholly shifted from the female to the male line, because society has never passed from mother-kin to father- kin. And in the Pelew Islands as in the ancient East we see the tide of political power running strongly in the direction of theocracy, the people resigning the conduct of affairs into the hands of men who claimed to rule them in the name of the gods. In the Pelew Islands such men might have developed into divine kings like those of Babylon and Egypt, if the natural course of evolution had not been cut short by the intervention of Europe.525 The evidence of the Khasis and the Pelew Islanders, two peoples very remote and very different from each other, suffices to prove that the influence which mother-kin may exert on religion is real and deep. But in order [pg 209] to dissipate misapprehensions, which appear to be rife on this subject, it may be well to remind or inform the reader that the ancient and widespread custom of tracing descent and inheriting property through the mother alone does not by any means imply that the government of the tribes which observe the custom is in the hands of women; in short, it should always be borne in mind that mother-kin does not mean mother-rule. On the contrary, the practice of mother-kin prevails most extensively amongst the lowest savages, with whom woman, instead of being the ruler of man, is always his drudge and often little better than his slave. Indeed, so far is the system from implying any social superiority of women that it probably took its rise from what we should regard as their deepest degradation, to wit, from a state of society in which the relations of the sexes were so loose and vague that children could not be fathered on any particular man.526 When we pass from the purely savage state to that higher plane of culture in which the accumulation of property, and especially of landed property, has become a powerful instrument of social and political influence, we naturally find that wherever the ancient preference for the female line of descent has been retained, it tends to increase the importance and enhance the dignity of woman; and her aggrandizement is most marked in princely families, where she either herself holds royal authority as well as private property, or at least transmits them both to her consort or her children. But this social advance of women has never been carried so far as to place men
- 25. f f as a whole in a position of political subordination to them. Even where the system of mother-kin in regard to descent and property has prevailed most fully, the actual government has generally, if not invariably, remained in the hands of men. Exceptions have no doubt occurred; women have occasionally arisen [pg 210] who by sheer force of character have swayed for a time the destinies of their people. But such exceptions are rare and their effects transitory; they do not affect the truth of the general rule that human society has been governed in the past and, human nature remaining the same, is likely to be governed in the future, mainly by masculine force and masculine intelligence. To this rule the Khasis, with their elaborate system of mother-kin, form no exception. For among them, while landed property is both transmitted through women and held by women alone, political power is transmitted indeed through women, but is held by men; in other words, the Khasi tribes are, with a single exception, governed by kings, not by queens. And even in the one tribe, which is nominally ruled by women, the real power is delegated by the reigning queen or High Priestess to her son, her nephew, or a more distant male relation. In all the other tribes the kingship may be held by a woman only on the failure of all male heirs in the female line.527 So far is mother-kin from implying mother-rule. A Khasi king inherits power in right of his mother, but he exercises it in his own. Similarly the Pelew Islanders, in spite of their system of mother-kin, are governed by chiefs, not by chieftainesses. It is true that there are chieftainesses, and that they indirectly exercise much influence; but their direct authority is limited to the affairs of women, especially to the administration of the women's clubs or associations, which [pg 211] answer to the clubs or associations of the men.528 And to take another example, the Melanesians, like the Khasis and the Pelew Islanders, have the system of mother-kin, being similarly divided into exogamous clans with descent in the female line; “but it must be understood that the mother is in no way the head of the family. The house of the family is the father's, the garden is his, the rule and government are his.”529 We may safely assume that the practice has been the same among all the many peoples who have retained the ancient system of mother-kin under a monarchical constitution. In Africa, for example, the chieftainship or kingship often descends in the female line, but it is men, not women, who inherit it.530 The theory of a gynaecocracy is in truth a dream of visionaries and pedants. And equally chimerical is the idea that the predominance of
- 26. f f goddesses under a system of mother-kin like that of the Khasis is a creation of the female mind. If women ever created gods, they would be more likely to give them masculine than feminine features. In point of fact the great religious ideals which have permanently impressed themselves on the world seem always to have been a product of the male imagination. Men make gods and women worship them. The combination of ancestor-worship with mother-kin furnishes a simple and sufficient explanation of the superiority of goddesses over gods in a state of society where these conditions prevail. Men naturally assign the first place in their devotions to the ancestress from whom they trace their descent. We need not resort to a fantastic hypothesis of the preponderance of the feminine fancy in order to account for the facts. The theory that under a system of mother-kin the women rule the men and set up goddesses for them to [pg 212] worship is indeed so improbable in itself, and so contrary to experience, that it scarcely deserves the serious attention which it appears to have received.531 But when we have brushed aside these cobwebs, as we must do, we are still left face to face with the solid fact of the wide prevalence of mother-kin, that is, of a social system which traces descent and transmits property through women and not through men. That a social system so widely spread and so deeply rooted should have affected the religion of the peoples who practise it, may reasonably be inferred, especially when we remember that in primitive communities the social relations of the gods commonly reflect the social relations of their worshippers. How the system of mother-kin may mould religious ideas and customs, creating goddesses and assigning at least a nominal superiority to priestesses over priests, is shown with perfect lucidity by the example of the Khasis, and hardly less clearly by the example of the Pelew Islanders. It cannot therefore be rash to hold that what the system has certainly done for these peoples, it may well have done for many more. But unfortunately through lack of documentary evidence we are seldom able to trace its influence so clearly.
- 27. § 3. Mother-Kin and Mother Goddesses in the Ancient East. While the combination of mother-kin in society with a preference for goddesses in religion is to be found as a matter of fact among the Khasis and Pelew Islanders of to-day, the former prevalence of mother-kin in the lands where the great goddesses Astarte and Cybele were worshipped is a matter of inference only. In later times father-kin had certainly displaced mother-kin among the Semitic worshippers of Astarte, and probably the same change had taken place among the Phrygian worshippers of Cybele. Yet the older [pg 213] custom lingered in Lycia down to the historical period;532 and we may conjecture that in former times it was widely spread through Asia Minor. The secluded situation and rugged mountains of Lycia favoured the survival of a native language and of native institutions long after these had disappeared from the wide plains and fertile valleys which lay on the highroads of war and commerce. Lycia was to Asia Minor what the highlands of Wales and of Scotland have been to Britain, the last entrenchments where the old race stood at bay. And even among the Semites of antiquity, though father-kin finally prevailed in matters of descent and property, traces of an older system of mother-kin, with its looser sexual relations, appear to have long survived in the sphere of religion. At all events one of the most learned and acute of Semitic scholars adduced what he regarded as evidence sufficient to prove “that in old Arabian religion gods and goddesses often occurred in pairs, the goddess being the greater, so that the god cannot be her Baal, that the goddess is often a mother without being a wife, and the god her son, and that the progress of things was towards changing goddesses into gods or lowering them beneath the male deity.”533 In Egypt the archaic system of mother-kin, with its preference for women over men in matters of property and inheritance, lasted down to Roman times, and it was traditionally [pg 214] based on the example of Isis, who had avenged her husband's murder and had continued to reign after his decease, conferring benefits on mankind. “For these reasons,” says Diodorus Siculus, “it was appointed that the queen should enjoy greater power and honour than the king, and that among private people the wife should rule
- 28. over her husband, in the marriage contract the husband agreeing to obey his wife in all things.”534 A corollary of the superior position thus conceded to women in Egypt was that the obligation of maintaining parents in their old age rested on the daughters, not on the sons, of the family.535 The same legal superiority of women over men accounts for the most remarkable feature in the social system of the ancient Egyptians, to wit, the marriage of full brothers with full sisters. That marriage, which to us seems strange and unnatural, was by no means a whim of the reigning Ptolemies; on the contrary, these Macedonian conquerors appear, with characteristic prudence, to have borrowed the custom from their Egyptian predecessors for the express purpose of conciliating native prejudice. In the eyes of the Egyptians “marriage between brother and sister was the best of marriages, and it acquired an ineffable degree of sanctity when the brother and sister who contracted it were themselves born of a brother and sister, who had in their turn also sprung from a union of the same sort.”536 Nor did the principle apply only to gods and kings. The common people acted on it in their daily life. They regarded marriages between brothers and sisters as the most natural and reasonable of all.537 The evidence of legal documents, [pg 215] including marriage contracts, tends to prove that such unions were the rule, not the exception, in ancient Egypt, and that they continued to form the majority of marriages long after the Romans had obtained a firm footing in the country. As we cannot suppose that Roman influence was used to promote a custom which must have been abhorrent to Roman instincts, we may safely assume that the proportion of brother and sister marriages in Egypt had been still greater in the days when the country was free.538 It would doubtless be a mistake to treat these marriages as a relic of savagery, as a survival of a tribal communism which knew no bar to the intercourse of the sexes. For such a theory would not explain why union with a sister was not only allowed, but preferred to all others. The true motive of that preference was most probably the wish of brothers to obtain for their own use the family property, which belonged of right to their sisters, and which otherwise they would have seen in the enjoyment of strangers, the husbands of their sisters. This is the system which in Ceylon is known as beena marriage. Under it the daughter, not the son, is the heir. She stays at home, and her husband comes and lives with her in the house; but her brother goes away and dwells in his wife's home, inheriting nothing from his parents.539 Such a system could not fail in time to prove irksome. Men would be loth to quit the old home, resign the ancestral property to a
- 29. stranger, and go out to seek their fortune empty-handed in the world. The remedy was obvious. A man had nothing to do but to marry his sister himself instead of handing her over to another. Having done so he stayed at home and enjoyed the family estate in virtue of his marriage with the heiress. This simple and perfectly effective expedient for keeping the property in the [pg 216] family most probably explains the custom of brother and sister marriage in Egypt.540 Thus the union of Osiris with his sister Isis was not a freak of the story- teller's fancy: it reflected a social custom which was itself based on practical considerations of the most solid kind. When we reflect that this practice of mother-kin as opposed to father-kin survived down to the latest times of antiquity, not in an obscure and barbarous tribe, but in a nation whose immemorial civilization was its glory and the wonder of the world, we may without being extravagant suppose that a similar practice formerly prevailed in Syria and Phrygia, and that it accounts for the superiority of the goddess over the god in the divine partnerships of Adonis and Astarte, of Attis and Cybele. But the ancient system both of society and of religion had undergone far more change in these countries than in Egypt, where to the last the main outlines of the old structure could be traced in the national institutions to which the Egyptians clung with a passionate, a fanatical devotion. Mother-kin, the divinity of kings and queens, a sense of the original connexion of the gods with nature—these things outlived the Persian, the Macedonian, the Roman conquest, and only perished under the more powerful solvent of Christianity. But the old order did not vanish at once with the official establishment of the new religion. In the age of Constantine the Greeks of Egypt still attributed the rise of the Nile to Serapis, the later form of Osiris, alleging [pg 217] that the inundation could not take place if the standard cubit, which was used to measure it, were not deposited according to custom in the temple of the god. The emperor ordered the cubit to be transferred to a church; and next year, to the general surprise, the river rose just as usual.541 Even at a later time Athanasius himself had to confess with sorrow and indignation that under his own eyes the Egyptians still annually mourned the death of Osiris.542 The end came with the destruction of the great Serapeum at Alexandria, the last stronghold of the heathen in Egypt. It perished in a furious and bloody sedition, in which Christians and pagans seem to have vied with each other in mutual atrocities. After its fall the temples were levelled with the ground
- 30. f f f or converted into churches, and the images of the old gods went to the melting-pot to be converted into base uses for the rabble of Alexandria.543 The singular tenacity with which the Egyptian people maintained their traditional beliefs and customs for thousands of years sprang no doubt from the stubborn conservatism of the national character. Yet that conservatism was itself in great measure an effect of geographical and climatic conditions and of the ways of life which they favoured. Surrounded on every side by deserts or almost harbourless seas, the Egyptians occupied a position of great natural strength which for long ages together protected them from invasion and allowed their native habits to set and harden, undisturbed by the subversive influence of foreign conquest. The wonderful regularity of nature in Egypt also conduced to a corresponding stability in the minds of the people. Year in, year out, the immutable succession of the seasons brought with it the same unvarying round of agricultural toil. What the fathers had done, the sons did in the same manner at the same season, and so it went on from [pg 218] generation to generation. This monotonous routine is common indeed to all purely agricultural communities, and everywhere tends to beget in the husbandman a settled phlegmatic habit of mind very different from the mobility, the alertness, the pliability of character which the hazards and uncertainties of commerce and the sea foster in the merchant and the sailor. The saturnine temperament of the farmer is as naturally averse to change as the more mercurial spirit of the trader and the seaman is predisposed to it. But the stereotyping of ideas and of customs was carried further in Egypt than in most lands devoted to husbandry by reason of the greater uniformity of the Egyptian seasons and the more complete isolation of the country. The general effect of these causes was to create a type of national character which presented many points of resemblance to that of the Chinese. In both we see the same inflexible strength of will, the same astonishing industry, the same strange blend of humanity and savagery, the same obstinate adherence to tradition, the same pride of race and of ancient civilization, the same contempt for foreigners as for upstarts and barbarians, the same patient outward submission to an alien rule combined with an unshakeable inward devotion to native ideals. It was this conservative temper of the people, bred in great measure of the physical nature of their land, which, so to say, embalmed the memory of Osiris long after the corresponding figures of Adonis and Attis had suffered decay. For while Egypt enjoyed profound repose, the tides of war and conquest, of traffic and commerce, had for
- 31. centuries rolled over Western Asia, the native home of Adonis and Attis; and if the shock of nationalities in this great meeting-ground of East and West was favourable to the rise of new faiths and new moralities, it was in the same measure unfavourable to the preservation of the old. [pg 219] Notes.
- 32. I. Moloch The King. I cannot leave the evidence for the sacred character of Jewish kings544 without mentioning a suggestion which was made to me by my friend and teacher the Rev. Professor R. H. Kennett. He thinks that Moloch, to whom first-born children were burnt by their parents in the valley of Hinnom, outside the walls of Jerusalem,545 may have been originally the human king regarded as an incarnate deity. Certainly the name of Moloch, or rather Molech (for so it is always written in the Massoretic text546 ), is merely a slightly disguised [pg 220] form of melech, the ordinary Hebrew word for “king,” the scribes having apparently given the dreadful word the vowels of bosheth, “shameful thing.”547 But it seems clear that in historical times the Jews who offered these sacrifices identified Molech, not with the human king, but with Jehovah, though the prophets protested against the custom as an outrage on the divine majesty.548 If, however, these sacrifices were originally offered to or in behalf of the human king, it is possible that they were intended to prolong his life and strengthen his hands for the performance of those magical functions which he was expected to discharge for the good of his people. The old kings of Sweden answered with their heads for the fertility of the ground,549 and we read that one of them, Aun or On by name, sacrificed nine of his sons to Odin at Upsala in order that his own life might be spared. After the sacrifice of his second son he received from the god an oracle that he should live as long as he gave him one of his sons every tenth year. When he had thus sacrificed seven sons, the ruthless father still lived, but was so feeble that he could no longer walk and had to be carried in a chair. Then he offered up his eighth son and lived ten years more, bedridden. After that he sacrificed his ninth son, and lived ten years more, drinking out of a horn like a weaned child. He now wished to sacrifice his last remaining son to Odin, but the Swedes would not let him, so he died and was buried in a mound at Upsala.550 In this Swedish tradition the king's children seem to have been looked upon as substitutes offered to the god in place of their father, and apparently this was also the current explanation of the slaughter of the first- born in the later times of Israel.551 On that view the sacrifices were
- 33. vicarious, and therefore purely religious, being intended to propitiate a stern and exacting deity. Similarly we read that when Amestris, wife of Xerxes, was grown old, she sacrificed on her behalf twice seven noble children to the [pg 221] earth god by burying them alive.552 If the story is true—and it rests on the authority of Herodotus, a nearly contemporary witness—we may surmise that the aged queen acted thus with an eye to the future rather than to the past; she hoped that the grim god of the nether-world would accept the young victims in her stead, and let her live for many years. The same idea of vicarious suffering comes out in a tradition told of a certain Hova king of Madagascar, who bore the sonorous name of Andriamasinavalona. When he had grown sickly and feeble, the oracle was consulted as to the best way of restoring him to health. “The following result was the consequence of the directions of the oracle. A speech was first delivered to the people, offering great honours and rewards to the family of any individual who would freely offer himself to be sacrificed, in order to the king's recovery. The people shuddered at the idea, and ran away in different directions. One man, however, presented himself for the purpose, and his offer was accepted. The sacrificer girded up his loins, sharpened his knife, and bound the victim. After which, he was laid down with his head towards the east, upon a mat spread for the purpose, according to the custom with animals on such occasions, when the priest appeared, to proceed with all solemnity in slaughtering the victim by cutting his throat. A quantity of red liquid, however, which had been prepared from a native dye, was spilled in the ceremony; and, to the amazement of those who looked on, blood seemed to be flowing all around. The man, as might be supposed, was unhurt; but the king rewarded him and his descendants with the perpetual privilege of exemption from capital punishment for any violation of the laws. The descendants of the man to this day form a particular class, called Tay maty manota, which may be translated, ‘Not dead, though transgressing.’ Instances frequently occur, of individuals of this class appropriating bullocks, rice, and other things belonging to the sovereign, as if they were their own, and escaping merely with a reprimand, while a common person would have to suffer death, or be reduced to slavery.”553 Sometimes, however, the practices intended to prolong the king's life seem to rest on a theory of nutrition rather than of substitution; in other words, the life of the victims, instead of being offered vicariously to a god, is apparently supposed to pass directly into the body of the sacrificer, thus
- 34. refreshing his failing strength and prolonging his existence. So regarded, the custom is magical rather than religious in character, since the desired effect is thought to follow directly without the intervention of a deity. At all events, it can be shown that sacrifices of this sort have been offered to prolong the life of kings in other parts of the world. Thus in regard to [pg 222] some of the negroes who inhabit the delta of the Niger we read that: “A custom which formerly was practised by the Ibani, and is still prevalent among all the interior tribes, consists in prolonging the life of a king or ancestral representative by the daily, or possibly weekly, sacrifice of a chicken and egg. Every morning, as soon as the patriarch has risen from his bed, the sacrificial articles are procured either by his mother, head wife, or eldest daughter, and given to the priest, who receives them on the open space in front of the house. When this has been reported to the patriarch, he comes outside and, sitting down, joins in the ceremony. Taking the chicken in his hand, the priest first of all touches the patriarch's face with it, and afterwards passes it over the whole of his body. He then cuts its throat and allows the blood to drop on the ground. Mixing the blood and the earth into a paste, he rubs it on the old man's forehead and breast, and this is not to be washed off under any circumstances until the evening. The chicken and the egg, also a piece of white cloth, are now tied on to a stick, which, if a stream is in the near vicinity, is planted in the ground at the water-side. During the carriage of these articles to the place in question, all the wives and many members of the household accompany the priest, invoking the deity as they go to prolong their father's life. This is done in the firm conviction that through the sacrifice of each chicken his life will be accordingly prolonged.”554 The ceremony thus described is, like so many other rites, a combination of magic and religion; for whereas the prayers to the god are religious, the passing of the victim over the king's body and the smearing of him with its blood are magical, being plainly intended to convey to him directly, without the mediation of any deity, the life of the fowl. In the following instances the practices for prolonging the king's life seem to be purely magical. Among the Zulus, at one of the annual feasts of first-fruits, a bull is killed by a particular regiment. In slaughtering the beast they may not use spears or sticks, but must break its neck or choke it with their bare hands. “It is then burned, and the strength of the bull is supposed to enter into the king, thereby prolonging his life.”555 Again, in an early Portuguese historian we read of a Caffre king of East Africa that “it is related of this Monomotapa
- 35. that he has a house where he commands bodies of men who have died at the hands of the law to be hung up, and where thus hanging all the humidity [pg 223] of their bodies falls into vases placed underneath, and when all has dropped from them and they shrink and dry up he commands them to be taken down and buried, and with the fat and moisture in the vases they say he makes ointments with which he anoints himself in order to enjoy long life—which is his belief—and also to be proof against receiving harm from sorcerers.”556 The Baganda of Central Africa used to kill men on various occasions for the purpose of prolonging the king's life; in all cases it would seem to be thought that the life of the murdered man was in some mysterious fashion transferred to the king, so that the monarch received thereby a fresh accession of vital energy. For example, whenever a particular royal drum had a new skin put on it, not only was a cow killed to furnish the skin and its blood run into the drum, but a man was beheaded and the spouting blood from the severed neck was allowed to gush into the drum, “so that, when the drum was beaten, it was supposed to add fresh life and vigour to the king from the life of the slain man.”557 Again, at the coronation of a new king, a royal chamberlain was chosen to take charge of the king's inner court and to guard his wives. From the royal presence the chamberlain was conducted, along with eight captives, to one of the human shambles; there he was blindfolded while seven of the men were clubbed to death, only the dull thud and crashing sound telling him of what was taking place. But when the seven had been thus despatched, the bandages were removed from the chamberlain's eyes and he witnessed the death of the eighth. As each man was killed, his belly was ripped open and his bowels pulled out and hung round the chamberlain's neck. These deaths were said to add to the King's vigour and to make the chamberlain strong and faithful.558 Nor were these the only human sacrifices offered at a king's coronation for the purpose of strengthening the new monarch. When the king had reigned two or three months, he was expected to hunt first a leopard and then a bushbuck. On the night after the hunt of the bushbuck, one of the ministers of State caught a man and brought him before the king in the dark; the king speared him slightly, then the man was strangled and the body thrown into a papyrus swamp, that it might never be found again. Another ceremony performed about this time to confirm the king in his kingdom was to catch a man, bind him, and bring him before the king, who wounded him slightly
- 36. with a spear. Then the man was put to death. These men were killed to invigorate the king.559 [pg 224] When a king of Uganda had reigned some time, apparently several years, a ceremony was performed for the sake of prolonging his life. For this purpose the king paid a visit—a fatal visit—to a chief of the Lung-fish clan, who bore the title of Nankere and resided in the district of Busiro, where the tombs and temples of the kings were situated. When the time for the ceremony had been appointed, the chief chose one of his own sons, who was to die that the king might live. If the chief had no son, a near relation was compelled to serve as a substitute. The hapless youth was fed and clothed and treated in all respects like a prince, and taken to live in a particular house near the place where the king was to lodge for the ceremony. When the destined victim had been feasted and guarded for a month, the king set out on his progress from the capital. On the way he stopped at the temple of the great god Mukasa; there he changed his garments, leaving behind him in the temple those which he had been wearing. Also he left behind him all his anklets, and did not put on any fresh ones, for he was shortly to receive new anklets of a remarkable kind. When the king arrived at his destination, the chief met him, and the two exchanged a gourd of beer. At this interview the king's mother was present to see her son for the last time; for from that moment the two were never allowed to look upon each other again. The chief addressed the king's mother informing her of this final separation; then turning to the king he said, “You are now of age; go and live longer than your forefathers.” Then the chief's son was introduced. The chief took him by the hand and presented him to the king, who passed him on to the body-guard; they led him outside and killed him by beating him with their clenched fists. The muscles from the back of the body of the murdered youth were removed and made into two anklets for the king, and a strip of skin cut from the corpse was made into a whip, which was kept in the royal enclosure for special feasts. The dead body was thrown on waste land and guarded against wild beasts, but not buried.560 When that ceremony was over, the king departed to go to another chief in Busiro; but on the way thither he stopped at a place called Baka and sat down under a great tree to play a game of spinning fruit-stones. It is a children's game, but it was no child's play to the man who ran to fetch the fruit-stones for the king to play with; for he was caught and speared to
- 37. death on the spot for the purpose of prolonging the king's life. After the game had been played the king with his train passed on and lodged with a certain princess till the anklets made from the muscles of the chief's murdered son were ready for him to wear; [pg 225] it was the princess who had to superintend the making of these royal ornaments.561 When all these ceremonies were over, the king made a great feast. At this feast a priest went about carrying under his mantle the whip that had been made from the skin of the murdered young man. As he passed through the crowd of merrymakers, he would flick a man here and there with the whip, and it was believed that the man on whom the lash lighted would be childless and might die, unless he made an offering of either nine or ninety cowrie shells to the priest who had struck him. Naturally he hastened to procure the shells and take them to the striker, who, on receiving them, struck the man on the shoulder with his hand, thus restoring to him the generative powers of which the blow of the whip had deprived him. At the end of the feast the drummers removed all the drums but one, which they left as if they had forgotten it. Somebody in the crowd would notice the apparent oversight and run after the drummers with the drum, saying, “You have left one behind.” The thanks he received was that he was caught and killed and the bones of his upper arm made into drumsticks for that particular drum. The drum was never afterwards brought out during the whole of the king's reign, but was kept covered up till the time came to bring it out on the corresponding feast of his successor. Yet from time to time the priest, who had flicked the revellers with the whip of human skin, would dress himself up in a mantle of cow-hide from neck to foot, and concealing the drumstick of human bones under his robe would go into the king's presence, and suddenly whipping out the bones from his bosom would brandish them in the king's face. Then he would as suddenly hide them again, but only to repeat the manoeuvre. After that he retired and restored the bones to their usual place. They were decorated with cowrie shells and little bells, which jingled as he shook them at the king.562 The precise meaning of these latter ceremonies is obscure; but we may suppose that just as the human blood poured into a drum was thought to pass into the king's veins in the booming notes of the drum, so the clicking of the human bones and the jingling of their bells were supposed to infuse into the royal person the vigour of the murdered man. The purpose of flicking commoners with the whip made of human skin is even more obscure; but we may conjecture that the life or virility of every man struck
- 38. f with the whip was supposed to be transmitted in some way to the king, who thus recruited his vital, and especially his reproductive, energies at this solemn feast. If I am right in my interpretation, all these Baganda [pg 226] modes of strengthening the king and prolonging his life belonged to the nutritive rather than to the vicarious type of sacrifice, from which it will follow that they were magical rather than religious in character. The same thing may perhaps be said of the wholesale massacres which used to be perpetrated when a king of Uganda was ill. At these times the priests informed the royal patient that persons marked by a certain physical peculiarity, such as a cast of the eye, a particular gait, or a distinctive colouring, must be put to death. Accordingly the king sent out his catchpoles, who waylaid such persons in the roads and dragged them to the royal enclosure, where they were kept until the tale of victims prescribed by the priest was complete. Before they were led away to one of the eight places of execution, which were regularly appointed for this purpose in different parts of the kingdom, the victims had to drink medicated beer with the king out of a special pot, in order that he might have power over their ghosts, lest they should afterwards come back to torment him. They were killed, sometimes by being speared to death, sometimes by being hacked to pieces, sometimes by being burned alive. Contrary to the usual custom of the Baganda, the bodies, or what remained of the bodies, of these unfortunates were always left unburied on the place of execution.563 In what way precisely the sick king was supposed to benefit by these massacres of his subjects does not appear, but we may surmise that somehow the victims were believed to give their lives for him or to him. Thus it is possible that in Israel also the sacrifices of children to Moloch were in like manner intended to prolong the life of the human king (melech) either by serving as substitutes for him or by recruiting his failing energies with their vigorous young life. But it is equally possible, and perhaps more probable, that the sacrifice of the first-born children was only a particular application of the ancient law which devoted to the deity the first-born of every womb, whether of cattle or of human beings.564 [pg 227]
- 40. II. The Widowed Flamen. § 1. The Pollution of Death. A different explanation of the rule which obliged the Flamen Dialis to resign the priesthood on the death of his wife565 has been suggested by my friend Dr. L. R. Farnell. He supposes that such a bereavement would render the Flamen ceremonially impure, and therefore unfit to hold office.566 It is true that the ceremonial pollution caused by death commonly disqualifies a man for the discharge of sacred functions, but as a rule the disqualification is only temporary and can be removed by seclusion and the observance of purificatory rites, the length of the seclusion and the nature of the purification varying with the degree of relationship in which the living stand to the dead. Thus, for example, if one of the sacred eunuchs at Hierapolis- Bambyce saw the dead body of a stranger, he was unclean for that day and might not enter the sanctuary of the goddess; but next day after purifying himself he was free to enter. But if the corpse happened to be that of a relation he was unclean for thirty days and had to shave his head before he might set foot within the holy precinct.567 Again, in the Greek island of Ceos persons who had offered the annual sacrifices to their departed friends were unclean for two days afterwards and might not enter a sanctuary; they had to purify themselves with water.568 Similarly no one might go into the shrine of Men Tyrannus for ten days after being in contact with the dead.569 Once more, at Stratonicea in Caria a chorus of thirty noble boys, clad in white and holding branches in their hands, used to sing a hymn daily in honour of Zeus and Hecate; but if one of them were sick or had suffered a domestic bereavement, he was for the time being excused, not permanently excluded, from the [pg 228] performance of his sacred duties.570 On the analogy of these and similar cases we should expect to find the widowed Flamen temporarily debarred from the exercise of his office, not permanently relieved of it. However, in support of Dr. Farnell's view I would cite an Indian parallel which was pointed out to me by Dr. W. H. R. Rivers. Among the Todas of the
- 41. Neilgherry Hills in Southern India the priestly dairyman (palol) is a sacred personage, and his life, like that of the Flamen Dialis, is hedged in by many taboos. Now when a death occurs in his clan, the dairyman may not attend any of the funeral ceremonies unless he gives up office, but he may be re- elected after the second funeral ceremonies have been completed. In the interval his place must be taken by a man of another clan. Some eighteen or nineteen years ago a man named Karkievan resigned the office of dairyman when his wife died, but two years later he was re-elected and has held office ever since. There have meantime been many deaths in his clan, but he has not attended a funeral, and has not therefore had to resign his post again. Apparently in old times a more stringent rule prevailed, and the dairyman was obliged to vacate office whenever a death occurred in his clan. For, according to tradition, the clan of Keadrol was divided into its two existing divisions for the express purpose of ensuring that there might still be men to undertake the office of dairyman when a death occurred in the clan, the men of the one division taking office whenever there was a death in the other.571 At first sight this case may seem exactly parallel to the case of the Flamen Dialis and the Flaminica on Dr. Farnell's theory; for here there can be no doubt whatever that it is the pollution of death which disqualifies the sacred dairyman from holding office, since, if he only avoids that pollution by not attending the funeral, he is allowed at the present day to retain his post. On this analogy we might suppose that it was not so much the death of his wife as the attendance at her funeral which compelled the Flamen Dialis to resign, especially as we know that he was expressly forbidden to touch a dead body or to enter the place where corpses were burned.572 But a closer inspection of the facts proves that the analogy breaks down at some important points. For though the Flamen Dialis was forbidden to touch a dead body or to enter a place where corpses were burned, he was permitted to attend a funeral;573 so that there could hardly be any objection to his attending the funeral of [pg 229] his wife. This permission clearly tells against the view that it was the mere pollution of death which obliged him to resign office when his wife died. Further, and this is a point of fundamental difference between the two cases, whereas the Flamen Dialis was bound to be married, and married too by a rite of special solemnity,574 there is no such obligation on the sacred dairyman of the Todas; indeed, if he is married, he is bound to live apart from his wife during his term of office.575 Surely the obligation laid on the Flamen Dialis to be married of
- 42. itself implies that with the death of his wife he necessarily ceased to hold office: there is no need to search for another reason in the pollution of death which, as I have just shown, does not seem to square with the permission granted to the Flamen to attend a funeral. That this is indeed the true explanation of the rule in question is strongly suggested by the further and apparently parallel rule which forbade the Flamen to divorce his wife; nothing but death might part them.576 Now the rule which enjoined that a Flamen must be married, and the rule which forbade him to divorce his wife, have obviously nothing to do with the pollution of death, yet they can hardly be separated from the other rule that with the death of his wife he vacated office. All three rules are explained in the most natural way on the hypothesis which I have adopted, namely, that this married priest and priestess had to perform in common certain rites which the husband could not perform without his wife. The same obvious solution of the problem was suggested long ago by Plutarch, who, after asking why the Flamen Dialis had to lay down office on the death of his wife, says, amongst other things, that “perhaps it is because she performs sacred rites along with him (for many of the rites may not be performed without the presence of a married woman), and to marry another wife immediately [pg 230] on the death of the first would hardly be possible or decent.”577 This simple explanation of the rule seems quite sufficient, and it would clearly hold good whether I am right or wrong in further supposing that the human husband and wife in this case represented a divine husband and wife, a god and goddess, to wit Jupiter and Juno, or rather Dianus (Janus) and Diana;578 and that supposition in its turn might still hold good even if I were wrong in further conjecturing that of this divine pair the goddess (Juno or rather Diana) was originally the more important partner. However it is to be explained, the Roman rule which forbade the Flamen Dialis to be a widower has its parallel among the Kotas, a tribe who, like the Todas, inhabit the Neilgherry Hills of Southern India. For the higher Kota priests are not allowed to be widowers; if a priest's wife dies while he is in office, his appointment lapses. At the same time priests “should avoid pollution, and may not attend a Toda or Badaga funeral, or approach the seclusion hut set apart for Kota women.”579 Jewish priests were specially permitted to contract the pollution of death for near relations, among whom father, mother, son, daughter, and unmarried sister are particularly enumerated; but they were forbidden to contract the pollution for strangers.
- 43. However, among the relations for whom a priest might thus defile himself a wife is not mentioned.580 § 2. The Marriage of the Roman Gods. The theory that the Flamen Dialis and his wife personated a divine couple, whether Jupiter and Juno or Dianus (Janus) and Diana, supposes a married relation between the god and goddess, and so far it would certainly be untenable if Dr. Farnell were right in assuming, on the authority of Mr. W. Warde Fowler, that the Roman gods were celibate.581 On that subject, however, Varro, the [pg 231] most learned of Roman antiquaries, was of a contrary opinion. He not only spoke particularly of Juno as the wife of Jupiter,582 but he also affirmed generally, in the most unambiguous language, that the old Roman gods were married, and in saying so he referred not to the religion of his own day, which had been modified by Greek influence, but to the religion of the ancient Romans, his ancestors.583 Seneca ridiculed the marriage of the Roman gods, citing as examples the marriages of Mars and Bellona, of Vulcan and Venus, of Neptune and Salacia, and adding sarcastically that some of the goddesses were spinsters or widows, such as Populonia, Fulgora, and Rumina, whose faded charms or unamiable character had failed to attract a suitor.584 Again, the learned Servius, whose commentary on Virgil is a gold mine of Roman religious lore, informs us that the pontiffs celebrated the marriage of the infernal deity Orcus with very great solemnity;585 and for this statement he would seem to have had the authority of the pontifical books themselves, for he refers to them in the same connexion only a few lines before. As it is in the highest degree unlikely that the pontiffs would solemnize any foreign rites, we may safely assume that the marriage of Orcus was not borrowed from Greek mythology, but was a genuine old Roman ceremony, and this is all the more probable because Servius, our authority for the custom, has recorded some curious and obviously ancient taboos which were observed at the marriage and in the ritual of Ceres, the goddess who seems to have been joined in wedlock to Orcus. One of these taboos forbade the use of wine, the other forbade persons to name their father or daughter.586 [pg 232]
- 44. Further, the learned Roman antiquary Aulus Gellius quotes from “the books of the priests of the Roman people” (the highest possible authority on the subject) and from “many ancient speeches” a list of old Roman deities, in which there seem to be at least five pairs of males and females.587 More than that he proves conclusively by quotations from Plautus, the annalist Cn. Gellius, and Licinius Imbrex that these old writers certainly regarded one at least of the pairs (Mars and Nerio) as husband and wife;588 and we have good ancient evidence for viewing in the same light three others of the pairs. Thus the old annalist and antiquarian L. Cincius Alimentus, who fought against Hannibal and was captured by him, affirmed in his work on the Roman calendar that Maia was the wife of Vulcan;589 and as there was a Flamen of Vulcan, who sacrificed to Maia on May Day,590 it is reasonable to suppose that he was assisted in the ceremony by a Flaminica, his wife, just as on my hypothesis the Flamen Dialis was assisted by his wife the Flaminica. Another old Roman historian, L. Calpurnius Piso, who wrote in the second century b.c., said that the name of Vulcan's wife was not Maia but [pg 233] Majestas.591 In saying so he may have intended to correct what he believed to be a mistake of his predecessor L. Cincius. Again, that Salacia was the wife of Neptune is perhaps implied by Varro,592 and is positively affirmed by Seneca, Augustine, and Servius.593 Again, Ennius appears to have regarded Hora as the wife of Quirinus, for in the first book of his Annals he declared his devotion to that divine pair.594 In fact, of the five pairs of male and female deities cited by Aulus Gellius from the priestly books and ancient speeches the only one as to which we have not independent evidence that it consisted of a husband and wife is Saturn and Lua; and in regard to Lua we know that she was spoken of as a mother,595 which renders it not improbable that she was also a wife. However, according to some very respectable authorities the wife of Saturn was not Lua, but Ops,596 so that we have two independent lines of proof that Saturn was supposed to be married. Lastly, the epithets “father” and “mother” which the Romans bestowed on many of their deities597 are most naturally understood [pg 234] to imply paternity and maternity; and if the implication is admitted, the inference appears to be inevitable that these divine beings were supposed to exercise sexual functions, whether in lawful marriage or in unlawful concubinage. As to Jupiter in particular his paternity is positively attested by Latin inscriptions, one of them very old, which describe Fortuna Primigenia, the great goddess of Praeneste, as his daughter.598 Again, the rustic deity
- 45. Welcome to our website – the perfect destination for book lovers and knowledge seekers. We believe that every book holds a new world, offering opportunities for learning, discovery, and personal growth. That’s why we are dedicated to bringing you a diverse collection of books, ranging from classic literature and specialized publications to self-development guides and children's books. More than just a book-buying platform, we strive to be a bridge connecting you with timeless cultural and intellectual values. With an elegant, user-friendly interface and a smart search system, you can quickly find the books that best suit your interests. Additionally, our special promotions and home delivery services help you save time and fully enjoy the joy of reading. Join us on a journey of knowledge exploration, passion nurturing, and personal growth every day! ebookbell.com