Professional Documents
Culture Documents
Full Chapter Adaptive and Adaptable Learning 11Th European Conference On Technology Enhanced Learning Ec Tel 2016 Lyon France September 13 16 2016 Proceedings 1St Edition Katrien Verbert PDF
Full Chapter Adaptive and Adaptable Learning 11Th European Conference On Technology Enhanced Learning Ec Tel 2016 Lyon France September 13 16 2016 Proceedings 1St Edition Katrien Verbert PDF
Full Chapter Adaptive and Adaptable Learning 11Th European Conference On Technology Enhanced Learning Ec Tel 2016 Lyon France September 13 16 2016 Proceedings 1St Edition Katrien Verbert PDF
https://textbookfull.com/product/collaboration-and-
technology-22nd-international-conference-criwg-2016-kanazawa-
japan-september-14-16-2016-proceedings-1st-edition-takaya-
eWork and eBusiness in Architecture Engineering and
Construction ECPPM 2016 Proceedings of the 11th
European Conference on Product and Process 2016
Limassol Cyprus 7 9 September 2016 1st Edition Symeon
Christodoulou
https://textbookfull.com/product/ework-and-ebusiness-in-
architecture-engineering-and-construction-ecppm-2016-proceedings-
of-the-11th-european-conference-on-product-and-
process-2016-limassol-cyprus-7-9-september-2016-1st-edition-
symeo/
Machine Learning and Knowledge Discovery in Databases
European Conference ECML PKDD 2016 Riva del Garda Italy
September 19 23 2016 Proceedings Part I 1st Edition
Paolo Frasconi
https://textbookfull.com/product/machine-learning-and-knowledge-
discovery-in-databases-european-conference-ecml-pkdd-2016-riva-
del-garda-italy-september-19-23-2016-proceedings-part-i-1st-
edition-paolo-frasconi/
https://textbookfull.com/product/software-architecture-13th-
european-conference-ecsa-2019-paris-france-
september-9-13-2019-proceedings-tomas-bures/
Adaptive and
Adaptable Learning
11th European Conference
on Technology Enhanced Learning, EC-TEL 2016
Lyon, France, September 13–16, 2016, Proceedings
123
Lecture Notes in Computer Science 9891
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen
Editorial Board
David Hutchison
Lancaster University, Lancaster, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Friedemann Mattern
ETH Zurich, Zürich, Switzerland
John C. Mitchell
Stanford University, Stanford, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
C. Pandu Rangan
Indian Institute of Technology, Madras, India
Bernhard Steffen
TU Dortmund University, Dortmund, Germany
Demetri Terzopoulos
University of California, Los Angeles, CA, USA
Doug Tygar
University of California, Berkeley, CA, USA
Gerhard Weikum
Max Planck Institute for Informatics, Saarbrücken, Germany
More information about this series at http://www.springer.com/series/7409
Katrien Verbert Mike Sharples
•
Adaptive and
Adaptable Learning
11th European Conference
on Technology Enhanced Learning, EC-TEL 2016
Lyon, France, September 13–16, 2016
Proceedings
123
Editors
Katrien Verbert Tomaž Klobučar
KU Leuven Jožef Stefan Institute
Leuven Ljubljana
Belgium Slovenia
Mike Sharples
The Open University
Milton Keynes
UK
LNCS Sublibrary: SL3 – Information Systems and Applications, incl. Internet/Web, and HCI
Chapters 2, 3, 17, 20, 22, 42, 45, 48, 56, 57, 76, 82, and 83 are distributed under the terms of the Creative
Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/). For further
details see license information in these chapters.
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, 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.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are
believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors
give a warranty, express or implied, with respect to the material contained herein or for any errors or
omissions that may have been made.
The dedicated work of all the PC members as well as the additional reviewers must
be acknowledged. Only with their help was it possible to deal with the high number of
submissions and still meet all deadlines as originally planned.
Keynote presentations completed this competitive scientific program. Pierre
Dillenbourg from the EPFL Center for Digital Education, Switzerland, gave a pre-
sentation on “How Does TEL Research Inform the Design of Educational Robots?”
and Vincent Aleven from Carnegie Mellon University presented on “Adaptivity in
Learning Technologies: Kinds, Effectiveness, and Authoring.” A keynote from the
European Commission covered policy aspects of technology enhanced learning.
A plenary panel session was held on the theme of the conference – Adaptive and
Adaptable Learning. Two invited panelists from the artificial intelligence and education
community, Benedict du Boulay and Rose Luckin, joined the researchers from the TEL
community.
Demonstrations and posters had a pronounced role in the conference program.
A plenary session was organized as a “TEL demo shootout” in which the demon-
strations were presented to arouse the audience’s curiosity and highlight the unique
aspects. Later on, the demonstrations were shown in action, giving participants the
opportunity for hands-on experience, sparking discussions between researchers, prac-
titioners, and educational developers, providing a basis to vote for the best demo.
A plenary session was dedicated to an exhibition of posters, to foster discussion about
work in progress and research issues. Representatives from the industry also presented
and discussed their contributions to the field in the industry track.
The TEL community proposed and organized a set of stimulating workshops as part
of the conference. In all, nine workshops were selected from the proposals and were
organized. Some of them continue a series of well-established workshops on motiva-
tional and affective aspects in TEL and on awareness and reflection in TEL. Others, like
Pedagogical Grounded Learning Analytics Design, were new for 2016. A doctoral
consortium was organized concurrently with the workshops, which provided an
opportunity for PhD students to discuss their work with experienced TEL researchers.
We would like to thank the many contributors for creating a stimulating conference
of high quality. These include foremost the authors, the PC members and reviewers,
and the conference chairs, who all contributed to the program. We would also like to
thank an enthusiastic and dedicated local organization team who made EC-TEL
Preface VII
a smooth and memorable experience. The conference was partially supported by the
European Association of Technology-Enhanced Learning (EATEL), Springer, and
EasyChair.
Executive Committee
General Chair
Tomaž Klobučar Jožef Stefan Institute, Slovenia
Program Chairs
Mike Sharples The Open University, UK
Katrien Verbert KU Leuven, Belgium
Workshop Chair
Christian Glahn University of Applied Sciences HTW Chur,
Switzerland
Industry Chairs
Franck Tarpin Bernard SBT Group, France
Pierre Dubuc OpenClassrooms, France
Dissemination Chairs
Christine Michel INSA de Lyon, France
Karim Sehaba Université Lumière Lyon 2, France
Program Committee
Additional Reviewers
Full Papers
Short Papers
Learning Analytics Pilot with Coach2 - Searching for Effective Mirroring . . . 363
Natasa Brouwer, Bert Bredeweg, Sander Latour, Alan Berg,
and Gerben van der Huizen
Demo Papers
Poster Papers
Finding the Needle in a Haystack: Who are the Most Central Authors
Within a Domain?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 632
Ionut Cristian Paraschiv, Mihai Dascalu, Danielle S. McNamara,
and Stefan Trausan-Matu
Learning Experiences Using Tablets with Children and People with Autism
Spectrum Disorder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 640
David Roldán-Álvarez, Ana Márquez-Fernández, Estefanía Martín,
and Cristian Guzmán
1 Introduction
Videos are considered one of the main resources for learning. For instance, YouTube
was ranked the second most popular social resource that has been used for informal
learning by students [1]. One of the challenges that faces the learners and tutors is the
tremendous amount of videos available in social environments (e.g. 300 h of video are
uploaded to YouTube every minute1). Finding the right videos can be time consuming,
especially if the learner is seeking knowledge in ill-defined domains such as culture or
body language.
Social interactions around videos (e.g. user’s textual comments, likes, dislikes, etc.)
offer a rich source of information about the video itself, the users, and the subject
domain. These interactions can provide access to diverse perspectives on the subject
domain and users can learn from each other vicariously.
In “The Wisdom of Crowds”, Surowiecki argues that one of the elements to have a
wise crowd is to have a diverse crowd [2]. A diverse crowd could provide different
1
http://www.statisticbrain.com/youtube-statistics/.
The rest of the paper is structured as follows: Sect. 2 positions this research in
related techniques used to analyse user comments and introduces a diversity framework
that informed the development of the model for this paper. Section 3 introduces the
proposed semantic-driven diversity model and the steps to operationalise the model.
Implementation of the model as the Semantic-Driven Diversity Analytic Tool (SeD-
DAT) is presented in Sect. 4. Section 5 shows the results from the application of
SeDDAT in a study with YouTube videos. Section 6 concludes and presents future
directions.
2 Related Work
Techniques for Classification and Ranking of Videos. Data mining techniques have
been used to exploit the richness of user interactions around videos, especially user
comments, for various purposes. For example, a mechanism for filtering comments was
proposed by Serbanoiu & Rebedea to identify relevant comments on YouTube videos
using classifications and ranking approaches [3]. Similarly, using classification tech-
niques a study by Siersdorfer et al. shows that community feedback on already rated
comments can help to filter and predict ratings for possibly useful and unrated com-
ments [4]. Using the state-of-the-art in learning to rank approaches, the user interac-
tions or “social features” were shown to be a promising approach for improving the
video retrieval performance in the work introduced by [5]. For improving video cat-
egorisation, a text-based approach was conducted to assign relevant categories to
videos, where the users’ comments among all the other features gave significant results
for predicting video categorisation [6]. Underpinned by data mining techniques,
Ammari et al. used user comments on YouTube videos to derive group profiles to
facilitate the design of learning simulated environments [7]. Galli et al. conducted a
study that used different data mining techniques to analyse user comments to introduce
a re-ranking method which produced a new ordered list of videos that is originally
provided by the YouTube recommender [8].
Semantics Techniques for Diversity Modelling. Semantics offers a great potential
for diversity modelling by providing an explicit structure to position the model within
the domain of interest. A new research stream in exploration of diversity of individual’s
views in social media platform has emerged. A formal framework has been developed
for extracting individual viewpoints from semantic tags associated with user comments
[9]. Research has shown that linked data can be a useful source for enriching user
modelling interactions when bringing new user dimensions, such as cultural variations
A Semantic-Driven Model for Ranking Digital Learning Objects 5
[10]. New work has also emerged on the interpretation and analysis of social web data
with a strong focus on cultural differences - for example, a comparison between Twitter
and Sina Weibo [11]. Likewise, recent work has also shown how data analytics can
benefit the workforce engagement in enterprise contexts [12].
Framework for Understanding Diversity. An extensive study by Andy Stirling on
measures for diversity shows how diversity has gained interest in different disciplines
such as ecology, economics and policy [13]. His study shows that diversity has been
measured based on three different dimensions, using Stilring’s terminology, variety,
balance and disparity. These dimensions have been used in three different ways to
indicate the level of diversity: one concept diversity (e.g. variety only as in ecology); or
dual concept diversity by combining two dimensions (e.g. variety and balance as used
in economics), or triple concept diversity as a combination of variety, balance and
disparity (e.g. as an aggregated value of the three dimensions as proposed by Stirling).
The Stirling framework has been used in different domains, such as cultural diversity
for policy and regulation [14], cultural diversity in the cinema, television and book
industries [15–17], and spread of subjects in interdisciplinary research [18]).
Informed by the Stirling diversity framework, this research uses the semantic
annotations of user comments on videos to facilitate video ranking according to
diversity.
The diversity dimensions, variety, balance and disparity are defined as follow
[19, p. 709]:
– Variety is “the number of categories into which system elements are apportioned”.
– Balance is “a function of the pattern of apportionment of elements across
categories”.
– Disparity is “the manner and degree in which the elements may be distinguished”.
Underpinned by semantic techniques, these dimensions will be used separately and
in combination as indicators to measure diversity in user comments against an ontology
representing a domain of interest, which will be labelled as domain diversity.
3.1 Preliminaries
Basic Components. The main input of the proposed model for measuring diversity is
a set of textual comments T ¼ ft1 ; t2 ; . . .; tn g which have been created by users U ¼
fu1 ; u2 ; . . .; um g while interacting with a set of digital objects D ¼ fd1 ; d2 ; . . .; dk g.
Social Cloud Components. Every digital object d has a set of users U ðdÞ ¼
fu1 ; u2 ; . . .; umd g who commented on d; where every user ui 2 U ðdÞ has written at
least one comment on d.
6 E. Abolkasim et al.
Ec ¼ f8e cjc 2 CX ^ Ec E
A Semantic-Driven Model for Ranking Digital Learning Objects 7
K ¼ f8cjjEc j [ 0g
v ¼ jK j ð1Þ
Balance b. The proportions pi of entities from annotation across the ontology super
classes that are identified for variety K. Shannon Entropy index is used for this
research. An alternative, Shannon Evenness, is not used as it will give infinity results
when variety is equal to one.
Xv
b¼ i¼1
pi ln pi ; where ð2Þ
jEc j
pi ¼
jcj
Disparity d. The manner and degree in which the entities from annotations may be
distinguished. This investigates how scattered/dispersed the entities from annotations
are within their super classes, which could be referred to as disparity within categories.
An internal validation index Ball-Hall [20], based on clustering, is adapted to measure
the dispersion dis(c) within each super class where a semantic distance measure
(shortest path [21]) is used to calculate the distances between entities for each super
class.
1 Xv
d¼ disðci Þ; where; ð3Þ
v i¼1
2
1 XjEc j
disðcÞ ¼ minðpathp ðej ; mc ÞÞ ;
jEc j j¼1 8p
2
A medoid is the most centrally located item in a cluster that has minimal average distances to all the
other items in the cluster [22].
8 E. Abolkasim et al.
Fig. 1. The process of producing ranked digital objects according to diversity in user comments.
digital object metadata. To calculate domain diversity, SeDDAT retrieves the entities
from an xml file and then uses the extracted entities for further calculations.
Output: Given the domain ontology, the algorithms of this tool calculate a vector of
the three diversity dimensions (variety, balance and disparity) for each digital object.
Figure 1 shows how SeDDAT is used for measuring the diversity in user comments.
The process goes through three layers: the social interactions layer, where the social
cloud (user comments, user profile, and digital object metadata) is collected; the
semantic layer, where a selected domain ontology is used to annotate the user com-
ments; and the diversity analytics layer, where SeDDAT extracts the entities used in the
annotations of the user comments, calculates the diversity of these entities that are
mapped against the domain ontology, and ranks the digital objects according to the
selected metrics.
In order to test the proposed diversity model, SeDDAT was used on a set of videos
about job interviews. Apart from the verbal communications, body language is one of
the aspects that may influence the outcome of the interaction between the interviewer
and interviewee. In an increasingly inclusive and diverse society, it is important to
understand the different possible interpretations of the body language signals to avoid
misunderstanding. This study aimed to test the usefulness of the diversity metrics in the
selection of videos that contain the most diverse range of comments relating to body
A Semantic-Driven Model for Ranking Digital Learning Objects 9
language in job interview. There is an assumption that the higher the diversity, the
higher the potential of a video for broadening and deepening the learners’ awareness.
Fig. 2. A protégé snapshot of the domain categories (top super classes) of the selected domain
ontology.
5.3 Results
The extracted entities from annotations were passed through the three algorithms
designed to calculate the diversity dimensions as shown in Fig. 1. The results (data
associated with each video as well as diversity dimensions) were saved in a spreadsheet
3
http://imash.leeds.ac.uk/ontology/amon/BodyLanguage.owl
10 E. Abolkasim et al.
Table 1. Sample results of seven YouTube videos sorted by video IDs (smallest to largest).
Video ID #Comments #Entities Variety Balance Disparity
103 25 6 2 0.32 39.4
190 5 2 1 0.01 60.5
209 74 48 4 0.68 20.08
363 4 16 3 0.39 25.28
402 425 105 6 1.14 10.65
403 293 68 4 0.85 14.83
788 45 35 5 0.75 15.95
for further analysis. Table 1 shows the diversity dimensions of a sample of seven
YouTube videos with some of the associated data: video ID and number of comments
and entities from annotations.
(1) Ranking Based on Variety. Videos with high variety indicate that the comments
have covered most or all of the high level aspects of the domain (i.e. the entities from
annotations are apportioned to different domain categories). Therefore, to identify
videos that covered a variety of domain aspects, the video ordering can be based on the
highest to smallest values for variety. As can be seen in Table 2, comments on the top
video 402 covered six domain categories (body sense function; body position; object;
body language; body motion; and body language signal meaning) compared to the
bottom ranking video 190 that had comments covering only one domain category
(body language signal meaning).
(2) Ranking Based on Balance. Videos with a high value in balance mean that
comments covered evenly the aspects of the domain (i.e. the entities from annotations
are well proportioned across domain categories). See Table 3 for the list of videos
sorted based on balance. The video 402 was ranked top, because the proportions pi of
its entities are higher compared to the other videos. Table 4 shows the proportions, as
defined in formula 2 in Sect. 3.2, of the two top videos 402 and 403. For example,
body language signal meaning has a total of 1336 entities (classes and instances), and
the proportions of videos 402 and 403 are 52 and 40 respectively.
A Semantic-Driven Model for Ranking Digital Learning Objects 11
Table 4. The proportions of videos 402 and 403 across the eight domain categories.
Video Body Body Body Body Object Nonverbal Kinesics Body
ID language position motion language (256) communication (1) senses
signal (33) (118) (429) (1) function
meaning (6)
(1336)
402 52 1 13 4 32 0 0 4
403 40 0 9 0 17 0 0 2
(3) Ranking Based on Disparity. Videos with high disparity indicate that the com-
ments cover distinctive aspects within the domain categories (i.e. the entities from
annotating the comments are widely scattered within their domain categories).
Therefore, to identify videos that triggered distinct domain aspects around their content,
the videos can be order largest to smallest according to their disparity value as can be
seen in Table 5.
Ranking based on disparity shifted the top videos (e.g. videos 402 and 403) that
were ranked based on variety or balance to the bottom. Similarly, the video 190 that
was ranked bottom for variety and balance came top here.
To investigate this further, the ranked videos were inspected closely using the
(a) video content, (b) number of comments, (c) number of entities from annotations,
and (d) samples of user comments. Also, a correlation between the number of user
comments and the diversity dimensions was conducted.
Another random document with
no related content on Scribd:
CHAPTER XXXI.
THE SUMMONS.
In the moonlight, bright as day, Mr. Jervis rode home beside Miss
Gordon’s rickshaw. Her tell-tale fan stuck out of the pocket of his
overcoat.
Yes, their little world was not blind; it was evidently a settled thing.
Most people were glad. The Brandes were sure to do the wedding in
“style;” and a wedding would be an agreeable variety from dances
and picnics.
“I shall come up to-morrow morning,” he said, as he reluctantly
released her hand, “to-morrow before twelve.”
Mr. Brande, who had effected his escape early, had returned
home, and been in bed and asleep for some hours.
He was suddenly aroused by his wife standing at his bedside, her
cloak hanging off her shoulders, her coiffeur a little deranged, a lamp
in her hand illuminating an unusually excited countenance.
“Well, what is it?” he demanded with pardonable irritation.
“Oh, P.! what do you think? A man has come from Simla——”
“Yes,” suddenly sitting erect, his official mind at once on the alert
for some pressing and important dispatch.
“He came out with them in the same ship,” she panted.
Had Sarabella his wife gone suddenly out of her mind?
“He says that Mark, not Waring, is the rich man.”
“He said it after supper, I suppose,” snarled Mr. Brande. “He was
drunk!”
“Not a bit of it! I tackled Mark himself, and he confessed. I was
very angry at being taken in. He declares they did it without meaning
a bit of harm at first, and that when it went too far he did not know
what to do. He is very sorry.”
“That he is a millionaire! Oh yes, I should think so!”
“He is coming up first thing to-morrow to tell you all about it; and,
unless I’m mistaken, to speak to you about Honor.”
“What about her?” sharply.
“Why, you dear, stupid man, are you asleep still? Can’t you
guess?”
“You told me that there was nothing of that sort; in fact,” with an
angry laugh, “that ‘the boy,’ as you called him, was desperately
devoted to you.”
“What stuff!” she ejaculated indignantly. “He will have thirty
thousand a year! I know that I shall never close an eye to-night!”
“And are good-naturedly resolved that I am to keep you in
countenance. You might, I think, have reserved this double-barrelled
forty-pounder for the morning.”
“And that’s all the thanks I get,” she grumbled, as she slowly
trailed away to her dressing-room.
Just about this very time, Mark Jervis was smoking a cigarette in
his bare sitting-room. Before him, on the table, lay a white feather
fan and a programme. He was much too happy to go to bed, he
wanted to sit up and think. His thoughts were the usual bright ones
incident to love’s young dream, and as he watched the smoke slowly
curling up the air was full of castles. These beautiful buildings were
somewhat rudely shattered by the entrance of his bearer—wrapped
in a resai, and looking extremely sleepy—with a letter in his hand.
“A Pahari brought this for the sahib three hours ago,” tendering a
remarkably soiled, maltreated envelope.
Of course it was from his father at last. He tore it open, and this
was what it said—
“My dear Son,
“I am very ill. If you would see me alive, come. The
messenger will guide you. I live forty miles out. Lose no time.
“Your affectionate father,
“H. Jervis.”
The letter was forty-eight hours old.
“Is the messenger here?” he asked eagerly.
“Yes, sahib.”
“Then call up the grey pony syce; tell him to take gram and a jule,
and saddle the pony. I am going off into the interior. I must start in
twenty minutes.”
The bearer blinked incredulously.
“I need not take you.” The bearer’s face expanded into a grin of
intense relief. “I shall be away several days. Get out my riding kit,
shove some clothes in a bag, and ask the cook to put up some bread
and meat and things, and tell the coolie I will be ready very shortly.”
Then he sat down, drew his writing-case towards him, and began
to write a note to Honor. Her first love letter—and strange, but true,
his also. It was merely a few lines to say he had been most suddenly
called away by his father, and hoped that he would be back within
the week.
It was both a keen disappointment and a keen pleasure to the girl
when the ayah brought the letter to her at nine o’clock. She read it
over and over again, but she will not allow our profane eyes to see it,
nor can it be stolen, for she carries it about with her by day, and it
rests under her pillow by night: at the end of the week it was getting
a little frayed.
When the ayah handed the note to the Miss Sahib, the writer was
already twenty miles out of Shirani, following a broad-shouldered
Gurwali with his head and shoulders wrapped in the invariable brown
blanket.
Their course was by mountain bridle-paths, and in an eastern
direction; the scenery was exquisite, but its beauties were entirely
lost upon Jervis, who was picturing other scenes in his mind’s eye.
The road crept along the sheer faces of bare precipices, or plunged
suddenly into woody gorges, or ran along a flat valley, with cultivated
fields and loosely built stone walls. The further they went, the lovelier
grew the country, the wilder the surroundings. At twelve o’clock they
halted to rest the grey pony—the messenger’s muscular brown legs
seemed capable of keeping up their long swinging trot all day. It was
four o’clock in the afternoon when they arrived at their journey’s end;
they abruptly descended into a flat wooded dale, surrounded by hills
on three sides, sloping away to the plains on the fourth. A path from
the bridle-road led them into a dense jungle of high grass, full of
cattle, pack ponies, and mules. Emerging from this, they came to a
wall, along which they kept for about three hundred yards, and
turning a sharp corner they found themselves outside a great square
yellow house, two stories high.
It seemed as if it had been bodily transplanted from England.
There was nothing irregular or picturesque about it—the windows
were in rows, the roof was square and had a parapet, the sole
innovation was a long verandah, which ran all round the building,
and was apparently of recent date, a mere after-thought.
Mark, as he rode up to the steps, looked about him for the coolie;
he had suddenly disappeared. There was no one to be seen. He
ascended to the verandah, it was deserted, save for some fowl, who
seemed delightfully at home. It was more the verandah of a native
dwelling than the entrance to the home of an Englishman.
The new-comer gazed around expectantly, and saw three string
charpoys, a bundle of dirty bedding, a pair of shoes, a huka, and a
turban.
The door, which was innocent of paint or bells, was ajar. He
pushed it open and found himself in a large, dim, very dirty hall. Here
he was confronted by an old nanny goat, and two kids; to the left he
saw a room, which appeared to be a mere repetition of the
verandah.
As he hesitated and looked about, a man suddenly appeared, a
servant presumably, wearing a huge red turban, and a comfortable
blue cloth coat. He was stout and well to do, had a fat face, a black
square beard, and remarkably thick lips.
He seemed considerably disconcerted, when he caught sight of
the stranger, but drawing himself up pronounced the words,
“Durwaza, Bund,” with overwhelming dignity. Adding in English—
“The sahib never see no one.”
“He will see me,” said Mark, with decision.
“Sahib sick, sar, seeing no one, those my orders. Sahib seeing no
sahibs for many years.”
“Well, he sent for me, and I have come. Let me see him
immediately. I am his son.”
The Mahomedan’s expression instantly changed from lofty
condescension to the most unqualified astonishment.
“The sahib’s—son!” he repeated incredulously.
“Yes. I have told you that once already. Look sharp, and send
some one to see after my pony; I have come a long distance.”
The bearer went away and remained absent about five minutes,
during which time Mark had leisure to note the dirt, and neglected,
almost ruinous, state of the house—which had originally been a fine
mansion—to listen to loud jabbering and whispering in the room
beside him, and to observe several pairs of native eyes eagerly
peeping through a crack in the door.
“Come with me,” said the bearer, with a sullen air. “The sahib will
see you presently.”
“Is he better?”
“Yes, he is quite well; please to sit here,” and he opened the door
of an immense dining-room, furnished with Bombay carved black
wood furniture, and a dusty Indian carpet. It was a room that was
evidently never used, and but rarely opened. Its three great long
windows, which were caked and dim with grime, looked out upon the
snows. This was evidently the back of the house; the front
commanded a view of the plains. The site had been admirably
selected.
A black tray, with cold meat and some very sour bad bread, was
borne in, and a place cleared on the dusty table by the joint efforts of
the sulky bearer and a khitmaghar, with a cast in his eye, and the
very leanest figure Mark had ever beheld. However, he was much
too hungry to be fastidious, and devoured the refreshments with a
capital appetite. Meanwhile, after their custom, the two men stood by
in silence with folded arms, staring with concentrated attention and
unremitting gaze until the conclusion of the meal.
It was quite dark when the bearer reappeared, and, throwing open
the door, announced in a deeply resentful tone—
“The sahib will see the sahib.”
Mark followed the fat, square, aggressive-looking back, till he
came to a curtained archway, and was ushered into a lofty dim room,
so dim, that he could barely discern the figure which rose to greet
him—a tall bent man in a dressing-gown.
“Mark, my boy, it was like you to come so soon,” said a shaky
voice. “Like what you were as a child,” and he held out both his
hands eagerly.
“I only got your letter at four o’clock this morning, sir,” said his son.
“I hope you are better?”
“I am for the present. I sent for you by a private messenger post-
haste, because I believed that I had but a few hours to live, and I
longed desperately to see you.”
“I have been hoping you would send for me for the last two
months. I have been waiting, as you know, in Shirani.”
“Yes—yes—yes! Sometimes the temptation was almost
irresistible, but I fought against it; for why should I cloud over your
young life? However, I had no choice; the situation has been forced
upon me—and you. My faithful companion, Osman, died ten days
ago, but we will talk of this another time. These voices in my head
interrupt me; especially that woman’s voice,” with an irritable
gesture.
His son could not, for the life of him, think of any immediate or
appropriate remark, and sat in embarrassed silence, and then Major
Jervis continued—
“You are six and twenty now—a grown man, Mark, and speak like
a man! I have not had a good look at your face yet. I wonder if it is
the same face as that of my own honest-eyed boy?”
The answer would be prompt, if he so pleased, for the lean
khitmaghar now staggered in under the weight of a large evil-
smelling “argand” lamp (a pattern extinct everywhere save in remote
parts of India).
Mark looked over eagerly at his father. His head was bent in his
hands. Presently he raised it, and gazed at his son with a look of
unmistakable apprehension. His son felt as if he were confronting an
utter stranger; he would never have recognized this grey-haired
cadaverous old man as the handsome stalwart sabreur he had
parted with sixteen years previously. He looked seventy years of
age. His features were sharpened as if by constant pain, his colour
was ashen, his hands emaciated, his eyes sunken; he wore a
camel’s-hair dressing-gown, and a pair of shabby slippers.
“You are just what I expected,” he exclaimed, after a long pause.
“You have your mother’s eyes; but you are a Jervis. Of course you
see a great change in me?”
“Well, yes—rather,” acquiesced his son, with reluctant truthfulness.
“India ages people.”
“You think this a strange life that I lead, I am sure; miles away from
my fellow-countrymen, buried alive, and long forgotten?”
“No, not forgotten, sir. Do you recollect Pelham Brande of the Civil
Service? He was asking for you only the other day.”
“I think I remember him—a clever fellow, with a very pretty wife,
who people said had been a servant. (How long these sort of things
stick to people’s memories.) I’ve been out of the world for years.”
“But you will return to it. Come back to England with me. What is
there to keep you in this country?”
“What, indeed!” with a jarring laugh. “No, my dear boy, I shall
never leave the Pela Bungalow, as they call it, until I am carried out
of it feet foremost.”
“Why do you say this? You are a comparatively young man—not
more than fifty-five.”
“I feel a thousand years old; and I often wish that I was dead.”
“I don’t wonder! I should say the same, if I had lived here alone for
seven years. How do you kill time?”
“I don’t kill time. Time is killing me. I walk in the garden sometimes,
but generally I sit and think. You must be tired, my boy,” as if struck
by a sudden thought.
“Well, I am, I must confess. I was at a ball until four o’clock this
morning.”
“A ball till four o’clock this morning!” he repeated. “How strange it
sounds. It seems the echo of a voice speaking twenty years ago!”
Dinner was served at a small table; a fowl for Mark, some patent
food for Major Jervis. The cooking was atrocious, the attendance
careless, the appointments splendid, but grimy. It was the same in
every department—an extraordinary mixture of squalor and
magnificence. It seemed to the indignant young man that these
ruffians of servants thought anything good enough for his father.
When Major Jervis’s huka was brought in he looked over at his
son and said—
“You smoke, of course?”
“Yes, thanks; but not that sort of thing. I would not know how to
work it.”
Last time he had lit a cigarette between four walls he little guessed
at the style of his next surroundings. The room was not
uncomfortable, the furniture was massively carved and luxurious, the
carpet rich Persian; there were book-cases full of volumes, and there
were fine pictures on the walls; but the paper was peeling off in
strips, and cobwebs hung like ropes from the corners. The books
were grimy with mould, the carpets and curtains inches deep in dust;
certainly a sort of oasis had been cleared around Major Jervis’s
chair, but everywhere the eye turned were tokens of neglect, poverty,
and decay. His father’s slippers were in holes, his linen frayed;
apparently he was a poor man. What had become of the begum’s
fortune?
CHAPTER XXXII.
“THE PELA KOTHI,” OR “YELLOW HOUSE.”