Deep Learning on Edge Computing Devices: Design

Challenges of Algorithm and Architecture 1st

Edition Xichuan Zhou - eBook PDF download

https://ebooksecure.com/download/deep-learning-on-edge-computing-
devices-design-challenges-of-algorithm-and-architecture-ebook-
pdf/

Download more ebook from https://ebooksecure.com

We believe these products will be a great fit for you. Click
the link to download now, or visit ebooksecure.com
to discover even more!

Artificial Intelligence and Machine Learning for EDGE

Computing 1st Edition Rajiv Pandey - eBook PDF

https://ebooksecure.com/download/artificial-intelligence-and-
machine-learning-for-edge-computing-ebook-pdf/

(eBook PDF) Deep Learning for Medical Image Analysis by

S. Kevin Zhou

http://ebooksecure.com/product/ebook-pdf-deep-learning-for-
medical-image-analysis-by-s-kevin-zhou/

(eBook PDF) Algorithm Design by Jon Kleinberg

http://ebooksecure.com/product/ebook-pdf-algorithm-design-by-jon-
kleinberg/

Fractional-order Design: Devices, Circuits, And Systems

1st Edition - eBook PDF

https://ebooksecure.com/download/fractional-order-design-devices-
circuits-and-systems-ebook-pdf/
Medical image recognition, segmentation and parsing :
machine learning and multiple object approaches 1st
Edition Zhou - eBook PDF

https://ebooksecure.com/download/medical-image-recognition-
segmentation-and-parsing-machine-learning-and-multiple-object-
approaches-ebook-pdf/

(Original PDF) Algorithm Design and Applications by

Michael T. Goodrich

http://ebooksecure.com/product/original-pdf-algorithm-design-and-
applications-by-michael-t-goodrich/

Artificial Intelligence and Deep Learning in Pathology

1st Edition Stanley Cohen Md (Editor) - eBook PDF

https://ebooksecure.com/download/artificial-intelligence-and-
deep-learning-in-pathology-ebook-pdf/

Deep Learning in Bioinformatics: Techniques and

Applications in Practice - eBook PDF

https://ebooksecure.com/download/deep-learning-in-bioinformatics-
techniques-and-applications-in-practice-ebook-pdf/

Circular Design for Zero Emission Architecture and

Building Practice 1st Edition - eBook PDF

https://ebooksecure.com/download/circular-design-for-zero-
emission-architecture-and-building-practice-ebook-pdf/
DEEP LEARNING
ON EDGE
COMPUTING
DEVICES
This page intentionally left blank
DEEP LEARNING
ON EDGE
COMPUTING
DEVICES
Design Challenges of Algorithm
and Architecture

XICHUAN ZHOU
HAIJUN LIU
CONG SHI
JI LIU
Elsevier
Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
Copyright © 2022 Tsinghua University Press. Published by Elsevier Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or any information storage and
retrieval system, without permission in writing from the publisher. Details on how to seek
permission, further information about the Publisher’s permissions policies and our arrangements
with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency,
can be found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the
Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience
broaden our understanding, changes in research methods, professional practices, or medical
treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described herein. In
using such information or methods they should be mindful of their own safety and the safety of
others, including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors,
assume any liability for any injury and/or damage to persons or property as a matter of products
liability, negligence or otherwise, or from any use or operation of any methods, products,
instructions, or ideas contained in the material herein.

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

ISBN: 978-0-323-85783-3

For information on all Elsevier publications

visit our website at https://www.elsevier.com/books-and-journals

Publisher: Mara Conner

Acquisitions Editor: Glyn Jones
Editorial Project Manager: Naomi Robertson
Production Project Manager: Selvaraj Raviraj
Designer: Christian J. Bilbow
Typeset by VTeX
Contents

Preface vii
Acknowledgements ix

PART 1 Introduction

1. Introduction 3
1.1. Background 3
1.2. Applications and trends 5
1.3. Concepts and taxonomy 8
1.4. Challenges and objectives 13
1.5. Outline of the book 14
References 16

2. The basics of deep learning 19

2.1. Feedforward neural networks 19
2.2. Deep neural networks 22
2.3. Learning objectives and training process 29
2.4. Computational complexity 33
References 34

PART 2 Model and algorithm

3. Model design and compression 39

3.1. Background and challenges 39
3.2. Design of lightweight neural networks 40
3.3. Model compression 47
References 56

4. Mix-precision model encoding and quantization 59

4.1. Background and challenges 59
4.2. Rate-distortion theory and sparse encoding 61
4.3. Bitwise bottleneck quantization methods 65
4.4. Application to efficient image classification 67
References 73

5. Model encoding of binary neural networks 75

5.1. Background and challenges 75

v
vi Contents

5.2. The basic of binary neural network 77

5.3. The cellular binary neural network with lateral connections 79
5.4. Application to efficient image classification 84
References 92

PART 3 Architecture optimization

6. Binary neural network computing architecture 97

6.1. Background and challenges 97
6.2. Ensemble binary neural computing model 98
6.3. Architecture design and optimization 102
6.4. Application of binary computing architecture 105
References 108

7. Algorithm and hardware codesign of sparse binary network

on-chip 111
7.1. Background and challenges 111
7.2. Algorithm design and optimization 115
7.3. Near-memory computing architecture 120
7.4. Applications of deep adaptive network on chip 124
References 135

8. Hardware architecture optimization for object tracking 139

8.1. Background and challenges 139
8.2. Algorithm 140
8.3. Hardware implementation and optimization 143
8.4. Application experiments 147
References 152

9. SensCamera: A learning-based smart camera prototype 155

9.1. Challenges beyond pattern recognition 155
9.2. Compressive convolutional network model 159
9.3. Hardware implementation and optimization 164
9.4. Applications of SensCamera 166
References 175

Index 179
Preface

We first started working in the field of edge computing-based machine

learning in 2010. With project funding, we tried to accelerate support
vector machine algorithms on integrated circuit chips to support embed-
ded applications such as fingerprint recognition. In recent years, with the
development of deep learning and integrated circuit technology, artificial
intelligence applications based on edge computing devices, such as intel-
ligent terminals, autonomous driving, and AIOT, are emerging one after
another. However, the realization of an embedded artificial intelligence ap-
plication involves multidisciplinary knowledge of mathematics, computing
science, computer architecture, and circuit and system design. Therefore
we arrived at the idea of writing a monograph focusing on the research
progress of relevant technologies, so as to facilitate the understanding and
learning of graduate students and engineers in related fields.
Deep learning application development based on embedded devices is
facing the theoretical bottleneck of high complexity of deep neural network
algorithms. Realizing the lightweight of various fast developing deep learn-
ing models is one of the keys to realize AIOT pervasive artificial intelligence
in the future. In recent years, we have been focusing on the development
of automated deep learning tools for embedded devices. This book covers
some of the cutting-edge technologies, currently developing in embedded
deep learning, and introduces some core algorithms, including lightweight
neural network design, model compression, model quantization, etc., aim-
ing to provide reference for the readers to design embedded deep learning
algorithm.
Deep learning application development based on embedded devices is
facing the technical challenge of limited development of integrated circuit
technology in the post-Moore era. To address this challenge, in this book,
we propose and elaborate a new paradigm of algorithm-hardware codesign
to realize the optimization of energy efficiency and performance of neural
network computing in embedded devices. The DANoC sparse coding neu-
ral network chip developed by us is taken as an example to introduce the
new technology of memory computing, hoping to give inspiration to em-
bedded design experts. We believe that, in the post-Moore era, the system
collaborative design method across multiple levels of algorithms, software,
and hardware will gradually become the mainstream of embedded intelli-
vii
viii Preface

gent design to meet the design requirements of high real-time performance

and low power consumption under the condition of limited hardware re-
sources.
Due to time constraints and the authors’ limited knowledge, there may
be some omissions in the content, and we apologize to the readers for this.

Xichuan Zhou
Acknowledgements

First of all, we would like to thank all the students who participated in the
relevant work for their contributions to this book, including Shuai Zhang,
Kui Liu, Rui Ding, Shengli Li, Songhong Liang, Yuran Hu, etc.
We would like to take the opportunity to thank our families, friends,
and colleagues for their support in the course of writing this monograph.
We would also like to thank our organization, School of Microelectron-
ics and Communication Engineering, Chongqing University, for providing
supportive conditions to do research on intelligence edge computing.
The main content of this book is compiled from a series of research,
partly supported by the National Natural Science Foundation of China
(Nos. 61971072 and 62001063).
We are most grateful to the editorial staff and artists at Elsevier and
Tsinghua University Press for giving us all the support and assistance needed
in the course of writing this book.

ix
This page intentionally left blank
PART 1

Introduction

1
This page intentionally left blank
CHAPTER 1

Introduction
1.1 Background
At present, the human society is rapidly entering the era of Internet of
Everything. The application of the Internet of Things based on the smart
embedded device is exploding. The report “The mobile economy 2020”
released by Global System for Mobile Communications Assembly (GSMA)
has shown that the total number of connected devices in the global Inter-
net of Things reached 12 billion in 2019 [1]. It is estimated that by 2025
the total scale of the connected devices in the global Internet of Things
will reach 24.6 billion. Applications such as smart terminals, smart voice
assistants, and smart driving will dramatically improve the organizational
efficiency of the human society and change people’s lives. With the rapid
development of artificial intelligence technology toward pervasive intelli-
gence, the smart terminal devices will further deeply penetrate the human
society.
Looking back at the development process of artificial intelligence, at
a key time point in 1936, British mathematician Alan Turing proposed
an ideal computer model, the general Turing machine, which provided
a theoretical basis for the ENIAC (Electronic Numerical Integrator And
Computer) born ten years later. During the same period, inspired by
the behavior of the human brain, American scientist John von Neumann
wrote the monograph “The Computer and the Brain” [2] and proposed
an improved stored program computer for ENIAC, i.e., Von Neumann
Architecture, which became a prototype for computers and even artificial
intelligence systems.
The earliest description of artificial intelligence can be traced back to
the Turing test [3] in 1950. Turing pointed out that “if a machine talks
with a person through a specific device without communication with the
outside, and the person cannot reliably tell that the talk object is a machine
or a person, this machine has humanoid intelligence”. The word “artificial
intelligence” actually appeared at the Dartmouth symposium held by John
McCarthy in 1956 [4]. The “father of artificial intelligence” defined it as
“the science and engineering of manufacturing smart machines”. The pro-
posal of artificial intelligence has opened up a new field. Since then, the
Deep Learning on Edge Computing Devices Copyright © 2022 Tsinghua University Press.
https://doi.org/10.1016/B978-0-32-385783-3.00008-9 Published by Elsevier Inc. All rights reserved. 3
4 Deep Learning on Edge Computing Devices

Figure 1.1 Relationship diagram of deep learning related research fields.

academia has also successively presented research results of artificial intel-

ligence. After several historical cycles of development, at present, artificial
intelligence has entered a new era of machine learning.
As shown in Fig. 1.1, machine learning is a subfield of theoretical re-
search on artificial intelligence, which has developed rapidly in recent years.
Arthur Samuel proposed the concept of machine learning in 1959 and con-
ceived the establishment of a theoretical method “to allow the computer
to learn and work autonomously without relying on certain coded instruc-
tions” [5]. A representative method in the field of machine learning is the
support vector machine [6] proposed by Russian statistician Vladimir Vap-
nik in 1995. As a data-driven method, the statistics-based SVM has perfect
theoretical support and excellent model generalization ability, and is widely
used in scenarios such as face recognition.
Artificial neural network (ANN) is one of the methods to realize
machine learning. The artificial neural network uses the structural and
functional features of the biological neural network to build mathematical
models for estimating or approximating functions. ANNs are computing
systems inspired by the biological neural networks that constitute animal
brains. An ANN is based on a collection of connected units or nodes
called artificial neurons, which loosely model the neurons in a biologi-
cal brain. The concept of the artificial neural network can be traced back
to the neuron model (MP model) [7] proposed by Warren McCulloch
and Walter Pitts in 1943. In this model the input multidimensional data
are multiplied by the corresponding weight parameters and accumulated,
 Introduction 5

and the accumulated value is calculated by a specific threshold function

to output the prediction result. Later, Frank Rosenblatt built a perceptron
system [8] with two layers of neurons in 1958, but the perceptron model
and its subsequent improvement methods had limitations in solving high-
dimensional nonlinear problems. Until 1986, Geoffrey Hinton, a professor
in the Department of Computer Science at the University of Toronto, in-
vented the back propagation algorithm [9] for parameter estimation of the
artificial neural network and realized the training of the multilayer neural
networks.
As a branch of the neural network technology, the deep learning tech-
nology has been a great success in recent years. The algorithmic milestone
appeared in 2006. Hinton invented the Boltzmann machine and success-
fully solved the problem [10] of vanishing gradients in training the mul-
tilayer neural networks. So far, the artificial neural network has officially
entered the “deep” era. In 2012, the convolutional neural network [11]
and its variants invented by Professor Yann LeCun from New York Uni-
versity greatly improved the classification accuracy of the machine learning
methods on large-scale image databases and reached and surpassed people’s
image recognition level in the following years, which laid the technical
foundation for the large-scale industrial application of the deep learning
technology. At present, the deep learning technology is ever developing
rapidly and achieved great success in subdivision fields of machine vision
[12] and voice processing [13]. Especially in 2016, Demis Hassabis’s Alpha
Go artificial intelligence built based on the deep learning technology de-
feated Shishi Li, the international Go champion by 4:1, which marked that
artificial intelligence has entered a new era of rapid development.

1.2 Applications and trends

The Internet of Things technology is considered to be one of the impor-
tant forces that lead to the next wave of industrial change. The concept of
the Internet of Things was first proposed by Kevin Ashton of MIT in 2009.
He pointed out that “the computer can observe and understand the world
by RF transmission and sensor technology, i.e., empower computers with
their own means of gathering information” [14]. After the massive data
collected by various sensors are connected to the network, the connection
between human beings and everything is enhanced, thereby expanding the
boundaries of the Internet and greatly increasing industrial production ef-
ficiency. In the new “wave of industrial technological change”, the smart
6 Deep Learning on Edge Computing Devices

terminal devices will undoubtedly play an important role. As a carrier for

connection of Internet of Things, the smart perception terminal device not
only realizes data collection, but also has front-end and local data process-
ing capabilities, which can realize the protection of data privacy and the
extraction and analysis of perceived semantic information.
With the proposal of the smart terminal technology, the fields of Arti-
ficial Intelligence (AI) and Internet of Things (IoT) have gradually merged
into the artificial intelligence Internet of Things (AI&IoT or AIoT). On
one hand, the application scale of artificial intelligence has been gradually
expanded and penetrated into more fields relying on the Internet of Things;
on the other hand, the devices of Internet of Things require the embedded
smart algorithms to extract valuable information in the front-end collection
of sensor data. The concept of intelligence Internet of Things (AIoT) was
proposed by the industrial community around 2018 [15], which aimed at
realizing the digitization and intelligence of all things based on the edge
computing of the Internet of Things terminal. AIoT-oriented smart ter-
minal applications have a period of rapid development. According to a
third-party report from iResearch, the total amount of AIoT financing in
the Chinese market from 2015 to 2019 was approximately $29 billion, with
an increase of 73%.
The first characteristic of AIoT smart terminal applications is the high
data volume because the edge has a large number of devices and large size
of data. Gartner’s report has shown that there are approximately 340,000
autonomous vehicles in the world in 2019, and it is expected that in 2023,
there will be more than 740,000 autonomous vehicles with data collec-
tion capabilities running in various application scenarios. Taking Tesla as
an example, with eight external cameras and one powerful system on chip
(SOC) [16], the autonomous vehicles can support end-to-end machine vi-
sion image processing to perceive road conditions, surrounding vehicles
and the environment. It is reported that a front camera with a resolution
of 1280 × 960 in Tesla Model 3 can generate about 473 GB of image data
in one minute. According to the statistics, at present, Tesla has collected
more than 1 million video data and labeled the information about dis-
tance, acceleration, and speed of 6 billion objects in the video. The data
amount is as high as 1.5 PB, which provides a good data basis for improve-
ment of the performance of the autonomous driving artificial intelligence
model.
The second characteristic of AIoT smart terminal applications is high la-
tency sensitivity. For example, the vehicle-mounted ADAS of autonomous
 Introduction 7

vehicles has strict requirements on response time from image acquisi-

tion and processing to decision making. For example, the average re-
sponse time of Tesla autopilot emergency brake system is 0.3 s (300 ms),
and a skilled driver also needs approximately 0.5 s to 1.5 s. With the
data-driven machine learning algorithms, the vehicle-mounted system
HW3 proposed by Tesla in 2019 processes 2300 frames per second (fps),
which is 21 times higher than the 110 fps image processing capacity of
HW2.5.
The third characteristic of AIoT smart terminal applications is high
energy efficiency. Because wearable smart devices and smart speakers in
embedded artificial intelligence application fields [17] are mainly battery-
driven, the power consumption and endurance are particularly critical.
Most of the smart speakers use a voice awakening mechanism, which can
realize conversion from the standby state to the working state according to
the recognition of human voice keywords. Based on the embedded voice
recognition artificial intelligence chip with high power efficiency, a novel
smart speaker can achieve wake-on-voice at standby power consumption
of 0.05 W. In typical offline human–machine voice interaction application
scenarios, the power consumption of the chip can also be controlled within
0.7 W, which provides conditions for battery-driven systems to work for a
long time. For example, Amazon smart speakers can achieve 8 hours of
battery endurance in the always listening mode, and the optimized smart
speakers can achieve up to 3 months of endurance.
From the perspective of future development trends, the development
goal of the artificial intelligence Internet of Things is achieving ubiquitous
pervasive intelligence [18]. The pervasive intelligence technology aims to
solve the core technical challenges of high volume, high time sensitivity,
and high power efficiency of the embedded smart devices and finally to
realize the digitization and intelligence of all things [19]. The basis of de-
velopment is to understand the legal and ethical relationship between the
efficiency improvement brought by the development of the artificial intel-
ligence technology and the protection of personal privacy, so as to improve
the efficiency of social production and the convenience of people’s lives
under the premise of guaranteeing the personal privacy. We believe that
pervasive intelligence calculation for the artificial intelligence Internet of
Things will become a key technology to promote a new wave of industrial
technological revolution.
8 Deep Learning on Edge Computing Devices

Figure 1.2 Global data growth forecast.

1.3 Concepts and taxonomy

1.3.1 Preliminary concepts
Data, computing power, and algorithms are regarded as three elements that
promote the development of artificial intelligence, and the development of
these three elements has become a booster for the explosion of the deep
learning technology. First of all, the ability to acquire data, especially large-
scale data with labels, is a prerequisite for the development of the deep
learning technology. According to the statistics, the size of the global Inter-
net data in 2020 has exceeded 30 ZB [20]. Without data optimization and
compression, the estimated storage cost alone will exceed RMB 6 trillion,
which is equivalent to the sum of GDP of Norway and Austria in 2020.
With the further development of the Internet of Things and 5G technol-
ogy, more data sources and capacity enhancements at the transmission level
will be brought. It is foreseeable that the total amount of data will continue
to develop rapidly at higher speed. It is estimated that the total amount of
data will be 175 ZB by 2025, as shown in Fig. 1.2. The increase in data
size provides a good foundation for the performance improvement of deep
learning models. On the other hand, the rapidly growing data size also puts
forward higher computing performance requirements for model training.
Secondly, the second element of the development of artificial intel-
ligence is the computing system. The computing system refers to the
hardware computing devices required to achieve an artificial intelligence
system. The computing system is sometimes described as the “engine” that
supports the application of artificial intelligence. In the deep learning era
of artificial intelligence, the computing system has become an infrastruc-
ture resource. When Google’s artificial intelligence Alpha Go [21] defeated
Korean chess player Shishi Li in 2016, people lamented the powerful artifi-
cial intelligence, and the huge “payment” behind it was little known: 1202
 Introduction 9

Figure 1.3 Development trend of transistor quantity.

CPUs, 176 high-performance GPUs, and the astonishing power of 233 kW

consumed in a game of chess.
From the perspective of the development of the computing system, the
development of VLSI chips is the fundamental power for the improve-
ment of AI computing performance. The good news is that although the
development of the semiconductor industry has periodic fluctuation, the
well-known “Moore’s law” [22] in the semiconductor industry has expe-
rienced the test for 50 years (Fig. 1.3). Moore’s law is still maintained in
the field of VLSI chips, largely because the rapid development of GPU has
made up for the slow development of CPU. We can see from the figure
that in 2010 the number of GPU transistors has grown more than that
of CPUs, CPU transistors have begun to lag behind Moore’s law, and the
development of hardware technologies [23] such as special ASICs for deep
learning and FPGA heterogeneous AI computing accelerators have injected
new fuel for the increase in artificial intelligence computing power.
Last but not least, the third element of artificial intelligence develop-
ment is an algorithm. An algorithm is a finite sequence of well-defined,
computer-implementable instructions, typically to solve a class of specific
problems in finite time. Performance breakthrough in the algorithm and
application based on deep learning in the past 10 years is an important rea-
son for the milestone development of AI technology. So, what is the future
development trend of deep learning algorithms in the era of Internet of Ev-
erything? This problem is one of the core problems discussed in academia
and industry. A general consensus is that the deep learning algorithms will
develop toward high efficiency.
10 Deep Learning on Edge Computing Devices

Figure 1.4 Comparison of computing power demands and algorithms for deep learn-
ing model.

OpenAI, an open artificial intelligence research organization, has

pointed out that “the computing resource required by advanced artificial
intelligence doubles approximately every three and a half months”. The
computing resource of training a large AI model has increased by 300,000
times since 2012, with an average annual increase of 11.5 times. The growth
rate of hardware computing performance has only reached an average an-
nual increase of 1.4 times. On the other hand, the improvement of the
efficiency of high-efficiency deep learning algorithms reaches annual aver-
age saving of about 1.7 times of the computing resource. This means that
as we continue to pursue the continuous improvement of algorithm per-
formance, the increase of computing resource demands potentially exceeds
the development speed of hardware computing performance, as shown in
Fig. 1.4. A practical example is the deep learning model GPT-3 [24] for
natural language processing issued in 2020. Only the cost of model training
and computing resource deployment has reached about 13 million dollars.
If the computing resource cost increases exponentially, then it is difficult
to achieve sustainable development. How to solve this problem is one of
the key problems in the development of artificial intelligence toward the
pervasive intelligence.

1.3.2 Two stages of deep learning: training and inference

Deep learning is generally classified into two stages, training and inference.
First, the process of estimating the parameters of the neural network model
based on known data is called training. Training is sometimes also known as
the process of parameter learning. In this book, to avoid ambiguity, we use
 Introduction 11

the word “training” to describe the parameter estimation process. The data
required in the training process is called a training dataset. The training al-
gorithm is usually described as an optimization task. The model parameters
with the smallest prediction error of the data labels on the training sample
set are estimated through gradient descent [25], and the neural network
model with better generalization is acquired through regularization [26].
In the second stage, the trained neural network model is deployed in the
system to predict the labels of the unknown data obtained by the sensor in
real time. This process is called the inference process. Training and inference
of models are like two sides of the same coin, which belong to different
stages and are closely related. The training quality of the model determines
the inference accuracy of the model.
For the convenience of understanding the subsequent content of this
book, we summarize the main concepts of machine learning involved in
the training and inference process as follows.
Dataset. The dataset is a collection of known data with similar at-
tributes or features and their labels. In deep learning, signals such as voices
and images acquired by the sensor are usually converted into data expres-
sion forms of vectors, matrices, or tensors. The dataset is usually classified
into a training dataset and a test dataset, which are used for the estimation
of the parameters of the neural network model and the evaluation of neural
network inference performance respectively.
Deep learning model. In this book, we will name a function f (x; θ)
from the known data x to the label y to be estimated as the model, where
θ is a collection of internal parameters of the neural network. It is worth
mentioning that in deep learning, the parameters and function forms of
the model are diverse and large in scale. It is usually difficult to write the
analytical form of the function. Only a formal definition is provided here.
Objective function. The process of deep learning model training is
defined as an optimization problem. The objective function of the op-
timization problem generally includes two parts, a loss function and a
regularization term. The loss function is used to describe the average error
of the label prediction of the neural network model on the training sam-
ples. The loss function is minimized to enhance the accuracy of the model
on the training sample set. The regularization term is usually used to con-
trol the complexity of the model to improve the accuracy of the model for
unknown data labels and the generalization performance of the model.
12 Deep Learning on Edge Computing Devices

Figure 1.5 Application scenarios of cloud and edge.

1.3.3 Cloud and edge devices

Edge computing [27] refers to a concept in which a distributed architec-
ture decomposes and cuts the large-scale computing of the central node
into smaller and easier-to-manage parts and disperses them to the edge
nodes for processing. The edge nodes are closer to the terminal devices and
have higher transmission speed and lower time delay. As shown in Fig. 1.5,
the cloud refers to the central servers far away from users. The users can
access these servers anytime and anywhere through the Internet to realize
information query and sharing. The edge refers to the base station or server
close to the user side. We can see from the figure that the terminal devices
[28] such as monitoring cameras, mobile phones, and smart watches are
closer to the edge. For deep learning applications, if the inference stage
can be completed at the edge, then the problem of transmission time de-
lay may be solved, and the edge computing provides services near data
sources or users, which will not cause the problem of privacy disclosure.
Data show that cloud computing power will grow linearly in future years,
with a compound annual growth rate of 4.6%, whereas demand at the edge
is exponential, with a compound annual growth rate of 32.5%.
The edge computing terminal refers to the smart devices that focus on
real-time, secure, and efficient specific scenario data analysis on user termi-
nals. The edge computing terminal has huge development prospects in the
field of artificial intelligence Internet of Things (AIoT). A large number
 Introduction 13

of sensor devices in the Internet of Things industry need to collect vari-

ous types of data at high frequency. Edge computing devices can integrate
data collection, calculation, and execution to effectively avoid the cost and
time delay of uploading the data to cloud computing and improve the se-
curity and privacy protection of user data. According to an IDC survey,
45% of the data generated by the Internet of Things industry in 2020 will
be processed at the edge of the network, and this proportion will expand
in the future years. “2021 Edge Computing Technology White Paper” has
pointed out that the typical application scenarios of edge computing smart
terminals include smart car networking/autonomous driving, industrial In-
ternet, and smart logistics. The values of ultralow time delay, massive data,
edge intelligence, data security, and cloud collaboration will prompt more
enterprises to choose edge computing.

1.4 Challenges and objectives

In recent years, deep learning has made breakthroughs in the fields of ma-
chine vision and voice recognition. However, because the training and
inference of standard deep neural networks involve a large number of pa-
rameters and floating-point computing, they usually need to be run on
resource-intensive cloud servers and devices. However, this solution has
the following two challenges.
(1) Privacy problem. Sending user data (such as photos and voice) to
the cloud will cause a serious privacy disclosure problem. The European
Union, the United States, etc. have set up strict legal management and
monitoring systems for sending the user data to the cloud.
(2) High delay. Many smart terminal applications have extremely high
requirements for the end-to-end delay from data collection to completion
of processing. However, the end-cloud collaborative architecture has the
problem that data transmission delay is uncertain and is difficult to meet
the needs of high time sensitivity smart applications such as autonomous
driving.
Edge computing effectively solves the above problem and has gradu-
ally become a research hotspot. Recently, edge computing has made some
breakthroughs in technology. On one hand, algorithm design companies
have begun to seek more efficient and lightweight deep learning models
(such as MobileNet and ShuffleNet). On the other hand, hardware tech-
nology companies, especially chip technology companies, have invested
heavily in the development of special neural network computing accel-
14 Deep Learning on Edge Computing Devices

eration chips (such as NPU). How to minimize resource consumption by

optimizing algorithms and hardware architecture on edge devices with lim-
ited resources is of great significance to the development and the application
of AIoT in the 5G and even 6G era.
The deep learning edge computing technology based on smart termi-
nals will effectively solve the above technical challenges of deep learning
cloud computing. This book focuses on the deep learning edge computing
technology and introduces how to design, optimize, and deploy efficient
neural network models on embedded smart terminals from the three lev-
els of algorithms, hardware, and applications. In the algorithm technology,
neural network algorithms for edge deep learning is introduced, includ-
ing lightweight neural network structure design, pruning, and compression
technology. The hardware technology details the hardware design and opti-
mization methods of edge deep learning, including algorithm and hardware
collaborative design, near memory computing, and hardware implementa-
tion of integrated learning. For the application program, each part briefly
introduces the application program. In addition, as a comprehensive ex-
ample, the application of smart monitoring cameras will be introduced as
a separate part at the end of this book, which integrates algorithm innova-
tion and hardware architecture innovation.

1.5 Outline of the book

This book aims to comprehensively cover the latest progress in edge-based
neural computing, including algorithm models and hardware design. To
reflect the needs of the market, in this book, we attempt to systematically
summarize the related technologies of edge deep learning, including algo-
rithm models, hardware architectures, and applications. The performance
of deep learning models can be maximized on the edge computing devices
through collaborative algorithm-hardware-code design.
The structure of this book is as follows. According to the content, it
includes three parts and nine chapters. Part 1 is Introduction, including
two chapters (Chapters 1–2); Part 2 is Model and Algorithm, including
three chapters (Chapters 3–5); and Part 3 is Architecture Optimization,
including four chapters (Chapters 6–9).
The first chapter (Introduction) mainly describes the development pro-
cess, related applications, and development prospects of artificial intelli-
gence, provides some basic concepts and terms in the field of deep learn-
 Another Random Document on
Scribd Without Any Related Topics
[510] Coke, The Complete Copyholder.
[511] Note-book of Bracton pl., 1237: “Dominus rex non vult se de
eis intromittere" (quoted Vinogradoff, Villainage in England, p. 46,
note 2).
[512] On this point see English Hist. Review, vol. viii. p. 296.
[513] Topographer and Genealogist, vol. i. The surveyor is
Humberstone.
[514] Calendar of Proceedings in Chancery in the Reign of Ed. VI.,
vol. i. p. cxxxvii.: “To the Right Honourable Sir Richard Riche, Kt.,
Lord Riche and Lord Chancellor of England. In humble wise
sheweth and complaineth unto your lordeshippe your daley orator
Richard Cullyer of Wymondham ... yeoman, and John Cullyer his
son,” that whereas they “were admitted tenants (of 20 acres) to
hold the same to them and their heirs ... and contynued seased of
the said 20 acres as of fee, as tenants at will, by copy of Court
Roll" now “Thomas Knyvett, Esq. ... of late claimed 10 acres of the
said 20 acres to be the demeanes of the said manor.” Knyvett (i.)
answers, “The said lond ys and have been tyme out of mynde
parcell of the demeanes of the moytie of the said manor of
Cromwell." (ii.) Denies that “the premises have been used to be
dymytted or be dymittable by copie of Court Roll for term of lyfe or
lyves as in fee"; on the contrary “yt may appear that the same
have been letten by term of yeres.”
[515] In 1548 an Act was passed “for the assurance to the tenants
of graunts and leases made for the Duke of Somerset’s demesne
lands.” It begins, “Whereas of truth noe custom or usage can or
maye by the lawes of this realm be annexed or knytt to any
meases, lands, tenements, or hereditaments letten by copye of
Court Roll ... albeyt those words 'secundum consuetudinem
manerii,' be rehearsed and expressed in the saide Court Rolle or
coppie had or made, except that the same meases, lands,
tenements, or other hereditaments, so letten be of olde customarie
or coppieholde land, and have byn used by all the tyme whereof
memory of man is not to the contrary to be letten or demysed by
copie of court roll."
[516] See pp. 122–123.
[517] Hist. MSS. Com., Cd. 3218, p. 74. Inquisition of February 20,
1308.
[518] Northumberland County History, vol. v. p. 282.
[519] Crondal Records (Baigent), Part I. p. 177.
[520] Northumberland County History, vol. viii.
[521] Acts of the Privy Council, vol. xiii. pp. 91–92, 1581. The
justices are to decide “if they thinke it agreeable with equite and
justice that the poore man should be put in possession of the said
landes."
[522] Croke’s Reports, vol. iii., Trin. 4 Caroli, Rot. dcciv. case 7.
Custom that copyholder for life may cut down trees pronounced “a
void and unreasonable custom and not allowable by law. For it is
the destruction of the inheritance and against the nature of a
copyholder for life. But peradventure there may be such a custom
for a copyholder of inheritance."
[523] Ibid., vol. iii., p. 198, case 8, Hill, 5 Car., Rot. 125: “The
question was whether a lord of a mannor may assess two years
and a half value of copyhold land according to racked rent for a
fine upon surrender and admittance, and for non-payment enter
for forfeiture. And all the Court conceived that one year and a half
of rent improved is high enough; and the defendant assessing two
years and a half it is unreasonable, and therefore the plaintiff
might well refuse the payment thereof." Ibid., vol. i. p. 779, case
13, takes the rule that unreasonable fines need not be paid back to
1600 ("It was holden per curiam that if the lord demands an
unreasonable fine of his coppyholder where the fine is uncertain, if
he denies it, it is not any forfeiture of his copyhold"), but his
judgment does not say how many years' rent is a reasonable fine.
The Calendar of Chancery Proceedings, temp. Eliz., is full of
petitions from tenants asking the court to declare fines excessive.
The rule that a fine must not exceed two years' rent does not
appear to have been accepted as binding till 1781 (Grant v. Ashe,
Douglas Reports, 722–723). But it is plain from the cases cited
above that by 1600 it was recognised that some fines were
unreasonable, and by 1630 that a reasonable fine should not
exceed one and a half years' rent. The fact that the Chancery
intervened to protect the equitable interests of copyholders earlier
than the Common Law Courts leads one to suspect that there must
be earlier cases than these of the Courts declaring fines
unreasonable. But I have not found them.
[524] Quarterly Journal of Economics, vol. xix.
[525] Roxburghe Club, Surveys of Pembroke Manors. The twenty-
one manors are as follows: Washerne, South Newton, North
Ugford, Brudecombe, Foughlestone, Chalke, Albedeston, Chilmerke
and Rugge, Staunton, Westoverton, Remesbury, Stockton,
Dichampton, Berwick St. John, Wyley, North Newton, Byshopeston
(all in Wilts), Donyett, Chedseye, South Brent (all in Somerset),
and Paynton in Devonshire. Estates of inheritance are found at
Byshopeston, and also fixed fines. At Paynton copies are granted
for 4 lives or less. The common formula for fines runs: “Pro talibus
finibus ut emptores et captores cum domino et officiariis suis
concordare vel barganizare possunt tam de terra in possessione
quam in reversione."
[526] MSS. Transcript in Wrexham Free Library by A.N. Palmer, of
“The Presentment and Verdict for the Manor of Hewlington,” 1620
(in which the proceedings in the reign of Elizabeth are recorded),
and “The Surveys of the Town and Liberty of Holt,” 1620. At
Hewlington it is stated that the Crown Commissioners made an
arrangement with the tenants “that if the said tenants would
relinquish these said pretended estates, revive the said decayed
rents, and pay two yeres Rent of the landes to the late Queen for a
fine, that then the said tenants and their heirs and assignes should
have leases granted them for fortie years, and so from fortie years
to fortie years in perpetuity.” It is not expressly stated that the
same arrangement was made at Holt, but it is to be inferred from
the context that it was.
[527] Coke, The Complete Copyholder.
[528] Northumberland County History, vol. viii. p. 238.
[529] See below, pp. 305–306.
[530] Northumberland County History, vol. viii. pp. 238–239.
[531] Ibid. For conditions on the Crown estates under Elizabeth
see S. P. D. Eliz., vol. xii. pp. 69–70: “Abstract of the Commission to
the lord Chancellor ... for letting the queen’s lands and tenements
in Northumberland within 20 miles of the border and in the
seigniories of Middleham and Richmond, Yorkshire and Barnard
Castle, Bishopric of Durham,” June 24, 1565.
[532] The manors are West Lexham (Holkham MSS., West Lexham,
No. 87, Map), Sparham (ibid., Sparham Bdle., No. 5), East
Dereham (R.O. Parliamentary Surveys, Norfolk, No. 1), Wighton
(R.O. Special Commissions, Duchy of Lancs., No. 839), Stockton
Socon (R.O. Parliamentary Surveys, Norfolk, No. 14), Aldeburgh
(R.O. Misc. Bks. Treas. of Receipts, vol. clxiii.), Chatesham (R.O.
ibid., vol. clxiii.), Dodnash (R.O. ibid., vol. clxiii.), Falkenham (R.O.
ibid., vol. clxiii.), Stratford iuxta Higham (R.O. Duchy of Lancaster,
Rentals and Surveys, 9/13), St. Edmund (R.O. Parliamentary
Surveys, Suffolk, No. 14), Mettingham (Victoria County History,
Suffolk), Mark Soham (ibid.).
[533] See Appendix II.
[534] E.g. Ormesby in Norfolk, where in 1516 thirty-one tenants
holding “in farm" formed the whole landholding population (R.O.
Rentals and Surveys, Gen. Ser., Portf. 22, No. 18). For a great rise
in rents following a probable substitution of leases for customary
tenures, see the case of Lewisham in Kent. On this manor in the
reign of Henry VI. the rent of the tenants (tenure unspecified) was
£8, 11s. 7d., 9 plougshares, and 6s. 2-1/2d. in the abbot’s hand
(R.O. Rentals and Surveys, Gen. Ser., Roll 361). In 1621 it was £23,
1s. 6-1/2d. (R.O. Misc. Bks. Treas. of Receipt, vol. clxxiv., fol. 1–
34). In the reign of James I. we have full details. The rent of the
free tenants was £17, 12s. 10-1/2d.; that of the tenants at will 9d.;
that of tenants “per dimissionem" (i.e. lease-holders) £72, 9s. 8-
1/2d. (R.O. Misc. Bks. Aug. Off., vol. ccccxiv., f. 33–34). It is
unfortunate that we are not told how the bulk of the tenants held
at the two earlier dates. But is it unreasonable to say that they
were probably customary tenants, and that the introduction of
leases was followed by a great rise in rents?
[535] Survey of Town and Liberty of Holt, MS. transcript in
Wrexham Free Library.
[536] S. P. D., ch. i. vol. cli., No. 38. (See Appendix I., No. iv.)
[537] Northumberland County History, vol. viii. p. 238.
[538] Ibid., vol. v. p. 211. The rent was raised from 18s. to 36s.
[539] Northumberland County History, vol. v. p. 210.
[540] Maitland, English Historical Review, vol, ix., “The History of a
Cambridgeshire Manor."
[541] Northumberland County History, vol. v.
[542] Roxburghe Club, Surveys of Pembroke Manors. The manors
are South Newton, Washerne, Donnington, Winterbourne Basset,
Estoverton and Phipheld, Byshopeston (all Wilts), and South Brent
and Huish (Somerset).
[543] E. E. T. S., Crowley, The Way to Wealth.
[544] See below, pp. 334–337.
[545] Calendar of Proceedings in Chancery in the Reign of Edward
VI. Bills to establish a fine certain on admission and alienation, to
get protection against exorbitant fines, &c. are common. For
popular complaints see E. E. T. S., A Supplication of the Poore
Commons: “These extortioners have so improved theyr lands that
they make of a xls. fyne xl. pounds,” &c. For an actual instance see
the following case. The tenants of Austenfield claim “that of
ancient time all the customary tenants of the said manor of
Austenfield were finable at fines certain, until of late years the
lords moved by covetousness, by troubling and vexing their
copyholders, drove many of them, for the buying of their
quietness, to be at fines uncertain" (William Salt Collection, vol. ix.
Chancery Proceedings. Bdle. 12, No. 70).
[546] Th. Wilson, A Discourse upon Usurie, 1584: “And therefore I
would not have men altogether to be enemies to the Canon Lawe,
and to condemn everything there written, because the Pope was
author of them.... Naie, I will saie plainlie that there be some such
lawes made by the Pope as be right godlie, saie others what they
list.”
[547] Norden, The Surveyor’s Dialogue, Book I.: “Surveyor. The
tennant leaveth commonly one either in right of inheritance, or by
surrender, to succeed him, and he by custome of the manor is to
be accepted tenant, alwaies provided he must agree with the lord,
if the custome of the manor hold not the fine certain as in few it
doth.... Farmer. You much mistake it, for I will show by ancient
court rolls that the fine of that which is now £20 was then but 13s.
4d., and yet will you say they are now as they were then?
Surveyor. Yea, and I thinke I erre little in it. For if you consider the
state of things then and now, you shall find the proportion little
differing; for so much are the prices of things vendible ... now
increased as may well be said to exceed the prices then as much
as £20 exceede the 13s. 4d."
 PART III
THE OUTCOME OF THE AGRARIAN
REVOLUTION

“Lords spiritual and temporal, have it in your

mind This
world as it waveth, and to your tenants be
kind.”
—The Proclamation of the Commons,
Gairdner,
Letters and Papers of Henry VIII.,
xii. I. 163.

“We must needs fight it out, or els be brought to the

lyke slavery that the Frenchmen are in.... Better yt were
therefore for us to dye like men, than after so great
misery in youth to dye more miserably in age."—E. E. T.
S., Crowley, The Way to Wealth.
Doctor. “On my faithe youe trouble youreselves ... youe
that be justices of everie countrie ... in sittinge upon
commissions almost wekely.”
Knight. “Surely it is so, yet the Kinge must be served
and the commonwealth. For God and the Kinge hathe
not sent us the poore lyving we have, but to doe
services therefore amonge our neighbours abroad."—
The Commonweal of this Realm of England.
“We have good Statutes made for the Commonwealth,
as touching commoners and inclosers, many meetings
and sessions; but in the end of the matter there cometh
nothing forth."—Latimer, First Sermon preached before
King Edward VI., March 8, 1549.
 CHAPTER I
THE AGRARIAN PROBLEM AND THE STATE

(a) The Political and Social Importance of the

PeasantryToC
The changes which have been described in the organisation of
agriculture created problems which were less absorbing than those
arising out of the religious reformation and the relation of England to
continental powers. When we turn over the elaborate economic
legislation of the reign of Elizabeth, with its attempts to promote
industry, to define class relationships, and to regulate with sublime
optimism almost every contract which one man can make with
another, we are tempted at first to see statesmen giving sleepless
nights to the solution of economic problems, and to think of a modern
bureaucratic state using the resources of scientific administration to
pursue a deliberate and clearly conceived economic policy. But this is
both to exaggerate the importance which economic questions
occupied in the minds of the governing aristocracies of the age, and
to credit them with a foresight which they did not possess. If they are
to be called mercantilists, in England, at any rate, they wear their
mercantilism with a difference; as a vague habit of mind, not as a
reasoned system of economic doctrines. Their administrative optimism
is the optimism of innocence as much as of omnipotence; the fruit of
a self-confidence which, in the name of the public interests, will prop
a falling trade, or cut down a flourishing one, with a bland naïveté
unperturbed by the hesitations which perplex even the most
courageous of modern protectionists. Though in several departments
of life—in commercial policy, in the regulation of the wage contract, in
the relief of distress—the main lines drawn by Elizabethan statesmen
will stand for two centuries, much of their legislation is very rough
and ready; much of it again is undertaken after generations of
dilatory experiments; much of it is devoid of any originality, and is a
mere reproduction on a national scale of the practice of individual
localities, a reproduction which sometimes does less than justice to
the original. If it is popular, it is popular because it tells men to do
what most decent men have been doing for a long time already, and
when it tells them to do something else it is carried out only with
great difficulty. If it is permanent, it is permanent not because
Parliamentary draughtsmen possess any great skill or foresight, but
because, before the rise of modern industry, all social relationships
have a great amount of permanence. Though there was much
interesting speculation on economic matters, economic rationalism
was as a practical force almost negligible; and since the only
instrument through which it could have achieved influence was the
monarchy, its lack of influence was perhaps politically fortunate.
Sixteenth century England was too busy getting the State on to its
feet to produce a Colbert. Lath and plaster Colberts built their card
castles on the Council table of James and Charles, and all was in train
for the sage paternal monarchy which was the ideal of Bacon. But a
wind blew from strange regions beyond their ken, and they were
scattered before they could do much either for good or evil, leaving,
as they fled, a cloud of dark suspicion round all those who would be
wiser in the art of Government than their neighbours, from which, in
the lapse of three centuries, the expert has hardly emerged. In spite
of mercantilism, economic questions never became in England the
pre-occupation of specialists. In spite of the genuine indignation
roused by the sufferings of the weaker classes in society, questions
affecting them were questions which statesmen did not handle for
their own sake, but only in so far as they forced themselves into the
circle of political interests by cutting across the order, or military
defence, or financial system, of the country. Apart from these high
matters of policy most members of the governing classes were
inclined to answer petitions on the subject of economic grievances as
Paget did to Somerset: Why can’t you let it alone? “What a good year
... is victuals so dear in England and nowhere else? If they and their
fathers before them have lived quietly these sixty years, pastures
being enclosed, the most part of these rufflers have least cause to
complain.”[548]
The subordinate place occupied by economic questions during our
period makes the attention which was given to the results of pasture-
farming all the more remarkable. Though to the statesmanship of the
sixteenth century the agrarian problem was one of the second order,
it was, at any rate till the accession of Elizabeth, the most serious of
its own class, and it was important enough to occupy Governments at
intervals for over a century and a half. The first Statute against
depopulation was passed in 1489;[549] an abortive Bill was introduced
into the House of Commons in 1656;[550] and between the two lies a
series of seven Royal Commissions, twelve Statutes, and a
considerable number of Proclamations dealing with one aspect or
another of the enclosing movement, as well as numerous decisions on
particular cases by the Privy Council, the Court of Star Chamber, and
the Court of Requests. This reaction of the new agrarian
developments upon public policy is interesting in several ways. It
illustrates the growth of new classes and forms of social organisation,
the methods and defects of sixteenth century administration, and the
ideas of the period as to the proper functions of the State in relation
to an important set of questions, upon which political opinion was in
some ways nearer to our own than it was to that of the age following
the Civil War. Nor, perhaps, is it altogether without importance from
the point of view of general history. We need not discuss how far the
reaction of some recent historians against the familiar judgments
which contrast Tudor tyranny with the constitutional revolutions of the
seventeenth century as darkness with light, is likely to be permanent.
But it is perhaps safe to say that it is in the sphere of social policy
that their case is seen at its strongest. After all, tyranny is often the
name which one class gives to the protection of another. To the small
copyholder or tenant farmer the merciless encroachments of his
immediate landlord were a more dreaded danger than the far-off
impersonal autocracy of the Crown to which he appealed for defence.
The period in which he suffered most in the sixteenth century was the
interval between the death of the despotic Henry VIII. and the
accession of the despotic Elizabeth. Though the interference of the
Tudor, and—in a feebler fashion—of the Stuart, Governments to
protect the peasantry was neither disinterested nor always effective,
its complete cessation after 1642, and the long line of Enclosure Acts
which follow the revolution of 1688, suggest that, as far as their
immediate economic interests were concerned, the smaller
landholders had more to lose than to gain from a revolution which
took power from the Crown to give it to the squires. The writers[551]
who after 1750 turned with a sigh from the decaying villages which
they saw around them, to glorify the policy of the absolutist
Governments of the sixteenth and seventeenth centuries, were
received with the ridicule which awaits all who set themselves against
a strong current of interests and ideas. But historically they were
right. The revolution, which brought constitutional liberty, brought no
power to control the aristocracy who, for a century and a half, alone
knew how such liberty could be used—that blind, selfish, indomitable,
aristocracy of county families, which made the British Empire and
ruined a considerable proportion of the English nation. From the
galleries of their great mansions and the walls of their old inns their
calm, proud faces, set off with an occasional drunkard, stare down on
us with the unshakable assurance of men who are untroubled by
regrets or perplexities, men who have deserved well of their order
and their descendants, and await with confidence an eternity where
preserves will be closer, family settlements stricter, dependents more
respectful, cards more reliable, than in this imperfect world they well
can be. Let them have their due. They opened a door which later
even they could not close. They fostered a tree which even they could
not cut down. But neither let us forget that to the poorer classes its
fruits were thorns and briars, loss of their little properties, loss of
economic independence, the hot fit of the hateful Speenhamland
policy, the cold fit of the more hateful workhouse system.[552] Those
who would understand the social forces of modern England must
realise that long disillusionment. Even in the seventeenth century
there are whisperings of it. At the end of the Civil War there were
men who were dimly conscious that the freedom for which they had
fought involved economic, as well as political and ecclesiastical,
changes. “Wee the poor impoverisht commoners,” wrote the leaders
of a little band of agrarian reformers to the Council of War in 1649,
“claim freedom in the common lands by vertue of this conquest over
the King, which is gotten by our joynt consent.... If this freedom be
not granted, wee that are the poor commoners are in a worse case
than we were in the King’s day.”[553] But from the reign of Henry VII.
to the Civil War official opinion was as generally in favour of
protecting the peasantry against the ruinous effects of agrarian
innovations, as it was on the side of leaving the landlords free to work
their will in the two centuries which succeeded. We must explain this
state of mind, for it certainly needs explanation; and this will
necessitate our looking at the movements of the peasants and at their
place in the State. We must estimate how far it was effective in
practice; and to do this we must say a few words about the
administrative machinery of the Tudors and of the first two Stuarts.
In almost all ages the first task of Governments is the preservation of
order. Though the economic ideas of the sixteenth century were very
different from those of the nineteenth, one of the reasons which
made it impossible for the statesmen of the period to leave the land
question altogether alone was the same as that which induced their
successors to deal with Irish land in 1870 and 1881. It was that
agrarian discontent created a permanent supply of inflammable
material, which a spark might turn into a conflagration. The years
between 1500 and 1650 are the last great age of the peasant
uprisings which, in all countries of Western Europe except France and
Ireland, are incredible to-day as a romance of giants, and hardly a
generation in that stormy period elapsed without one. Sometimes
nothing more happened than a collision of justices and gentry with
angry mobs who were tearing down hedges and restoring common to
common again under mysterious figures who flit across the darkening
country-side with weapons in their hands and the eternal insurrection
of the New Testament on their lips—Jack o' the Style, Pyrce Plowman,
and that prophetic Captain Pouch, who “was sent of God to satisfie all
degrees whatsoever, and in this present work was directed by the
Lord of Heaven.”[554] Sometimes the discontent swelled to a small
civil war, as it did in Lincolnshire and Yorkshire in 1536, and in the
eastern and southern counties in 1549. The Lincolnshire rising and
the Pilgrimage of Grace were, it is true, mainly motived by discontent
with the attack on the abbeys. But the explanation of their objects
given by those insurgents who were cross-examined by the
Government makes it difficult to agree with Professor Gay that only
an insignificant part was played in these movements by agrarian
discontent. The truth is that we ought to distinguish between the
objects of different sections. The rebels of 1536 were not a class, but
almost the whole society of northern England, which suddenly rolls
forward with all its members, spirituality and laity, peasants and
peers, in fervent motion together. The weaker side of these great
conservative demonstrations was that, though all classes were united
against the régime typified by Cromwell, all classes were not moved
to the same degree by the same grievances. Even when the old
religion was the cause that took the gentry into the field, the humbler
rebels were brought out as much by hatred of agrarian as of religious
innovations. The men of Lincolnshire marched under a banner
embroidered with a ploughshare, and laggards were spurred forward
with the cry “What will ye do? Shall we go home and keep
sheep?”[555] In Cumberland the four Captains of Penrith—Faith,
Poverty, Pity, and Charity—marched in solemn procession with drawn
swords round Burgh Church, and then, having heard Mass, led their
followers, with the blessing of the vicar, on a crusade to put an end to
gentlemen and to withhold rents and fines.[556] In the North generally
the arrival of Aske’s messengers was a signal for the wholesale
plucking down of new enclosures; a programme of agrarian reform
was included in the demands put forward at Doncaster; and Aske
himself told the Government at his examination that the practice of
letting out farms over the heads of poor tenants was one of the
causes of the rising.[557] A well-informed officer of State like Sir
William Paget seems to have thought that even the rebellion which
took place in Devonshire and Somersetshire in 1549, the causes of
which were mainly ecclesiastical, was partly also agrarian.[558] In that
year, indeed, nearly the whole of the southern counties, beginning in
May with Hertfordshire, from Norfolk in the east to Hampshire in the
south and Worcester in the west, were driven into riot by
disappointment with the ineffective Royal Commission appointed in
the preceding year. In 1550 there were disturbances in Kent, and the
Government anticipated their appearance in Essex. In 1552 the
Buckinghamshire peasants rose on account of high rents and high
prices. In 1554 Wyatt's[559] adherents demanded that all pasture
lands which had forcibly been seized by persons in power should be
restored. In 1569 an armed band pulled down enclosures near
Chinley[560] in Derbyshire, threatened to kill the encloser, and rescued
by force those of their number who were arrested. Twenty-six years
later, at a time of unusually high prices, even the peasantry of
Oxfordshire,[561] that most imperturbable of English counties, planned
“to knock down the gentlemen and rich men who made corn so dear,
and who took the commons.” In 1607 in the Midlands, where in the
preceding decade enclosure and depopulation had created a situation
as acute as that of half a century before, there was a riot which
resulted in the appointment of a Royal Commission.
This was perhaps the last serious agrarian rising which England has
seen. But though henceforward the hatred of the new agrarian
régime ran for the most part underground, it had been burned too
deep into the minds of the people to be lightly forgotten, and more
than once its smouldering embers flickered up in occasional riots. In
the first flush of the army’s victory over King and Parliament, when
the shattering of authority seemed for a moment to make all things
new, not only the political, but the economic, ideas of two centuries
later burst for a moment, as in an early spring, into wonderful and
premature life. The programme of the Levellers, who more than any
other party could claim to express the aspirations of the unprivileged
classes, included a demand not only for annual or biennial
Parliaments, manhood suffrage, a redistribution of seats in proportion
to population, and the abolition of the Veto of the House of Lords, but
also “that you would have laid open all enclosures of fens and other
commons, or have them enclosed only or chiefly for the benefit of the
poor.”[562] Theoretical communism, repudiated by some of the
Levellers, found its expression in the agitation of the Diggers, those
“true born sons and friends of England" who, under Everard and
Winstanley, set themselves, in the spirit of an Owenite Community, to
convert the waste land at Weybridge into the New Jerusalem.[563] For
to many earnest souls the day of the Lord seems very near, and Israel
must make ready against it, not with anguish of spirit only, but with
spade labour upon the barren earth. The contrast between the
prevalence of organised agrarian revolts in the middle of the sixteenth
century, dragging on in small sporadic agitations for nearly one
hundred years, with their comparative rarity two hundred years later,
when similar causes were at work to produce them, marks the new
grouping of social classes and economic forces which was going on
apace in our period. The intelligence of toiling England, that for a
century now has gone to build up a new civilisation in factory and
mine, in trade union and co-operative store, still lay in the larger
villages, its immemorial home. Discontent travelled across the
enclosing counties as it does to-day in a Welsh mining valley,
outcoursing oppression itself, like Elijah running before Ahab into
Jezreel. “If three or four good fellows would ride in the night with
every man a bell, and cry in every town that they pass, 'To Swaffham!
To Swaffham!' by the morning there would be ten thousand
assembled at the least; and then one bold fellow to stand forth and
say, 'Sirs, now we be here assembled, you know how little favour the
gentlemen bear us poor men.... Let us ... harness ourselves.'”[564]
Good fellows and bold were not wanting. “From that time forward no
man could keep his servant at plough; but every man that could bear
a staff went forward.”[565] Before the appearance of almost universal
leasehold tenure, standing armies, and omnipotent aristocratic
Parliaments, unrest among the rural population might cause the
Government a not inexpensive campaign, in which the reluctant
militia of yesterday were the enthusiastic rebels of to-day, and there
was not therefore much disparity between the discipline and
equipment of the forces engaged on either side. Both in the mainly
agrarian revolts in Norfolk, and in the mainly religious revolts in
Devonshire, the peasants fell, as they hoped they might, like men,
and it was the arquebuses of the foreign mercenaries which really
decided the struggle. Poor homeless hirelings, what could they know
but to clamour for their pay, and shoot better men than themselves?
To understand the nature of a body at rest it is sometimes advisable
to look at the same body when it is in motion. The agrarian
disturbances of our period possess certain features which are of
interest even to those who are concerned primarily not with social
politics, but with economic organisation. In the first place, they mark
the transition from the feudal revolts of the fifteenth century, based
on the union of all classes in a locality against the central
government, to those in which one class stands against another
through the opposition of economic interests. In the Lincolnshire
rebellion and in the Pilgrimage of Grace the old spirit predominated.
In the North of England the new agrarian régime had not proceeded
far enough to sap entirely the ancient bonds between landlord and
tenant, and the plunder of the monastic estates had not yet set a
commercial aristocracy in the seat of the old-fashioned Catholic
landlords. The commons of Westmoreland, who declare that they will
trust no gentlemen with their councils, nevertheless feel sufficient
confidence in Lord Darcy to write to him for his advice as to how far
they will be justified in insisting on reduced admission fines, and in
pulling down “all the intakes yt be noysum for poor men.”[566] Had
the Catholic gentry generally been willing to sacrifice the rents got
from pasture-farming, these movements might have found leaders
who would have made them more formidable. As it was, even when
hatred of the religious changes or of some particular piece of
legislation, like the unpopular Statute of Uses, enrolled the gentry
with the peasants, as in Lincolnshire and Yorkshire in 1536, the
incompatibility of the allies was obvious, and the presence of the
wealthier classes inspired distrust among the rank and file, who saw
in them the authors of their economic evils, and who, though
genuinely concerned at the painful destruction of the social
institutions of the old religion, were fighting mainly for the
maintenance of “old customs and tenant right,” fair rents and security
of tenure. In spite of the temporary union of all classes in 1536, the
insurgents tended to break up into two camps corresponding roughly
with the division between landlord and tenant. In Lincolnshire, though
the commons were influenced by the gentry so far as to demand the
repeal of the Act of Uses, “not knowing,” as a witness said, “what that
Act of Uses meant,” they showed their distrust of the upper classes by
refusing to allow them to discuss their future policy apart from the
general body of insurgents, while the extremists clamoured that “they
ought to kill some of the justices; also that if they hanged for this,
they would not leave one gentleman alive in Lincolnshire.”[567] At
Richmond all lords and gentlemen were to swear on the mass-book to
maintain the profit of Holy Church, to take nothing of their tenants
but the usual rents, to put down Cromwell and not to go to London,
on pain of death if they refused.[568] For courts have strange arts of
seduction, and though London (thank Heaven) is not England now, it
was still less England then. The rough rhymes that ran through the
North contain the warning of all popular movements against the
treachery of leaders, the sad eternal warning which buoys the sands
where so many high endeavours have gone to wreck. “All commons
stick ye together, rise with no great man till ye know his intent. Keep
your harness in your hands, and ye shall obtain all your purpose in all
this North land.... Claim ye old customs and tenant right, to take your
farms by a God's penny, all gressums and heightenings to be laid
down. Then may we serve our sovereign Lord King Henry VIII., God
save his noble Grace.
We shall serve our lands’ lords in every righteous cause
With horse and harness as custom will demand.
Lords spiritual and temporal have it in your mind
This world as it waveth, and to your tenants be kind.
Adieu, gentle commons, thus make I an end:
Writer of this letter, pray Jesu be his speed;
He shall be your captain, when that ye have need.”[569]
The temporary solidarity which had drawn all classes into the
Pilgrimage of Grace, though it flickered up for the last time in the
feudal revolt of the northern earls in 1569, was absent altogether
from the widespread agitation of 1549 to 1550. Except in Devonshire
and Cornwall, the disturbances of those years were purely agrarian, a
movement of tenants against landlords. The Eastern rebels were for
leaving “as many gentlemen in Norfolk as there be white bulls";[570]
the gentry responded by rallying to the Government; and both in that
country and in Devonshire the military forces which put down the
peasants were led by the two most notoriously unpopular landlords in
England, who had built up their estates out of confiscated abbey
lands, the Earl of Warwick and Sir William Herbert. In the reign of
Henry VII. the problem before Governments had still been to prevent
a great landlord from using his authority over his tenants to make war
on his neighbours or on the State. Sixty years later it is to prevent
tenants in several different counties from combining against landlords.
The landed classes recognise the new spirit. They denounce the
peasants as communists and agitators; and when they get a free
hand, as in the years from 1549 to 1553, they insist on legislation
which will make effective combination impossible.
In the second place, the way in which the agrarian agitations were
conducted is interesting as showing both the comparative prosperity
and independence of the English peasantry, even at a time when the
fortunes of many of them were declining, and the general conceptions
of social expediency held by what was regarded as the most
representative part of the English nation. It would be a mistake to
think of the rebels who joined these revolts as mere unorganised
malcontents, with nothing to lose. There is no resemblance at all,
either in personnel or methods, between the agrarian disturbances of
our period and the riots of starving agricultural labourers who burned
ricks under Captain Swing in the early nineteenth century. The
peasants who formed the backbone of the movements were often
well-to-do men, who were fighting to keep their land with the
dreadful tenacity of small proprietors. They had arms and were
accustomed to their use. They had sufficient money to raise common
funds. They included among their number sanguine and pertinacious
litigants who, so far from being disposed to throw up their case at the
hint of the landlord’s displeasure, were quite capable of making his
life one long lawsuit. The readiness of a class to make effective the
protection given it by the law in the face of the opposition of powerful
individuals, quenched, alas! too often by ignorance, and timidity, and
generations of dull oppression, is a very good test of its spirit and of
the practical freedom which it enjoys. In the sixteenth century,
though we certainly see many gross cases of intimidation, we also see
tenants appealing to the law courts and to the Government over the
heads of lords of manors. Such appeals are a proof of the
helplessness of the victims which has been commented on above. But
they are also a proof of the persistence and cohesion of some among
them. For while in the absence of oppression they would not have
been necessary, in the absence of a determination to resist
oppression they could not have been made. To enclose was in parts of
the country to stir up a hornet’s nest. There was not much
obsequiousness about the villagers of Thingden,[571] who from 1494
to 1538 pursued their landlord through almost every Court in the
Kingdom. The leaders of the popular agitation were often the more
prosperous among the middle-classes. Sanders, the general in the
interminable struggle over the common lands of the city of Coventry
which began in 1460, was a member of the important craft of Dyers,
and had occupied the high civic office of Chamberlain.[572] At
Louth[573] the initiative among the commons was taken by a tailor
and a weaver. Ket[574] himself was a considerable landed proprietor
as well as a tanner.
The peasants' agitations took the form both of more or less organised
risings and of sporadic rioting, which aimed at ends varying from
place to place according to the grievances inspired by the varying
conditions of different districts. Everywhere there were the throwing
down of enclosures and the driving of sheep.[575] In Yorkshire the
enclosures which were pulled down seem to have been mainly intakes
from the waste, and in Norfolk and the Midlands enclosures of arable
land which had been converted to pasture. In Warwickshire the Earl
of Warwick’s park was demolished, while in Wiltshire, where Sir
William Herbert had acquired the lands of Wilton Abbey, and enclosed
a whole village in his new park at Washerne, the peasants rose and
tore down the palings.[576] In the North generally the bitterest outcry
seems to have arisen over the excessive fines and “gressums”
charged for the admission of copyholders. In Cumberland[577] there
was a general strike against the payment of rents, and almost
everywhere there were complaints of the diminution in the area
available for pasturing the beasts of commoners through the
enclosing by landlords of manorial wastes.
Though it involves abandoning the order of events, let us illustrate by
a single example[578] the shape assumed by agrarian rioting, which
has not yet become a rebellion. In the summer of 1569, when Cecil
and Elizabeth were waiting anxiously for news from those northern
counties which “know no other prince but a Percy,” there was much
running and riding, much sending for warrants and plentiful delay in
their execution, in the wild country between Chinley and Bakewell,
whose centre is the Peak, and whose principal gorge now carries the
most beautiful piece of railway line in England. The Derbyshire
peasantry seem to have been ill to deal with. A few years later some
of those in Glossopdale succeeded in setting the Earl of Shrewsbury
at defiance, and, when evicted from their farms, induced the Council
to intervene to insist on their reinstatement.[579] Just now those of
them who lived in the neighbourhood of Chinley were in a ferment
over the enclosure of some common land. The story is a curious one,
and shows both the kind of conditions under which agrarian
discontent developed, and the way in which it was associated in the
mind of the Government with fears of political disturbance. The Duchy
of Lancaster, to whom the land near Chinley belonged, had let a
parcel of herbage called Mayston Field to one Lawrence Wynter, his
lease to begin as soon as that of the existing tenant had expired. In
that age of land speculation land changed hands rapidly. On the same
day as Wynter obtained the lease he sold it to a certain Richard Celey.
Celey transferred it to Godfrey Bradshaw, and Godfrey Bradshaw got
rid of it to his brother Anthony. The trouble began when the land
came into the hands of Godfrey Bradshaw. He started to hedge and
ditch it, which of course involved the exclusion of the other
inhabitants from the rights of pasture which they had hitherto
enjoyed. Accordingly the villagers, led by twelve of their number, of
whom four belonged to one family, removed the ditch, tore down the
enclosure, which consisted of “XLIII hundredth quicksetts willowes
and willowe stackes ... and did utterlye destroy and cutt the sayd
stacks and quick setts in pieces,” proceeding at the same time, with
the object of protecting their own grazing land against
encroachments, themselves to divide up the land into smaller
enclosures to be held by each man in severalty. Godfrey Bradshaw
then obtained warrants for the preservation of the peace against the
ringleaders, and at the same time induced the lessor, who was Sir
Ralph Sadler, the Chancellor of the Duchy, to address a letter to them
directing them not to interfere with any houses, hedges, or ditches,
which might in future be constructed round the land. They received
his communication, but massed in force with arms on Chinley Hill,
pulled down what still remained of Bradshaw’s hedges, and then
proceeded to organise the nucleus of a very pretty agitation. They
gave part of the herbage, which was nominally in the occupation of
the unfortunate lessee, to one William Beard, on condition that, after
the manner of his betters in the good old days before the Tudors, he
should “maynteyn them geynst the Queenes Majestie,” his support
taking the form of an agreement that he “should from tyme to tyme
send them Ydill ryotouse p'sons to assyste them in these yll doinges.”
They then raised a fund, presumably by a levy on the inhabitants,
called a meeting in the forest of High Peak, and set off about the
tenth of June to Bakewell for a further conference, arranging in the
meantime that some one should burn Godfrey Bradshaw’s house, and
that while his enclosures, if re-erected, should be pulled down, the
other inhabitants should make haste to divide up the disputed land
into twenty-one separate parcels. When the Bradshaws, having got
their warrants, tried with the aid of the village constable to execute
them, their opponents (“the land was grabbed from him, and he did
what any decent man would do”[580]) threatened them with murder,
and, on one of the party being actually arrested, came very near to
carrying their threat out. “The said p'tyes ... did ryotouslye assemble
themselves together in great companies at the town of Hayfield with
unlawfull weapons, that is to saye, with bowes, pytchefforkes,
clobbes, staves, swords, and daggers drawen, and ryotouslye dyd
then and there assaulte and p'sue the sayd Godfrey and Edward
Bradshawe, and in ryotouse manner dyd reskewe and take from them
the body of the sayd Richard Shower, being attached; the Queenes
Officer, George Yeavely of Bawdon, then being p'sent commanding
the peace to be kepte.” Having chased the enemy for some distance,
they camped on the contested territory, and kept a watchful eye and
a firm hand for any sign of the reappearance of the detested hedges.
More serious still in the eyes of the Government (and this, one
suspects, was their undoing), the leaders of this village revolution
went so far as to entangle themselves in high politics. At their
examination they are asked, “Whether dyd Reynold Kirke about May
day last paste, and dyvers tymes since and before, or any other tyme,
confederate, consulte, practise, or otherwise confer and talk with one
Mr. Bircles of the countye of Chester ... touching or concerning
prophesis by noblemen, or otherwise, and what books of prophesie
have you or the said Bircles seen or heard, and what is the effect
thereof, and how often have you or he perused, used, or conferred of
the same, or about such purposes, and with whom?" We do not know
how they answered this question. It may be that the anger of these
Derbyshire peasants at their vanishing commons was indeed a
fraction to be set among weightier assets by schemers in high places,
and that the sinister Mr. Bircles had really talked with them of matters
more serious than the pulling down of hedges and the baiting of
enclosers, of things forbidden to the vulgar, of the scattering of
upstart officials, of the restoration of a Catholic monarchy, of Mary,
who in the previous year had made her irrevocable plunge across the
Border. It may be merely that all in authority had that autumn an
unusually bad attack of nerves. In 1569 the North was full of
prophets, both noble and other.
It was not always the case, however, that agrarian discontent ended
in casual rioting of this kind. Of mere destructive violence there is,
indeed, in all the social disturbances of the period, singularly little.
There was a good deal in the routine of rural life, with its common
administration of land and dependence on a collectively binding
custom, to teach habits of discipline and co-operation. It must be
remembered that those who took the initiative in breaking the law
were not the peasants who pulled down enclosures, but the landlords
who made them in defiance of repeated statutes forbidding them. On
the whole the organised character of the action taken is more
conspicuous than the individual excesses, and if one is to look for a
modern analogy to the mixture of deliberation and violence which it
shows, it must be sought in an Irish fair rent campaign rather than in
the bread riots of a despairing urban proletariat. When the agitation
was confined to individual manors it occasionally took the form of
agrarian trade unionism. Tenants collectively decline to serve as jurors
in the court of the manor till their demands are granted.[581] They
raise a common purse.[582] They refuse to pay more than a certain
rent. When more than one manor is implicated different localities
display a rough cohesion. Whole communities seem to have joined
the movement in 1536 and 1540 with a certain formality. In
Lincolnshire and Yorkshire townships were brought out on the ringing
of the town bell with the cohesion of a well-organised trade union;
Beverley[583] sent messages to the Lincolnshire rebels under its
common seal; and the part which was played by the village officers in
the movements of the peasantry is proved by the Proclamation[584]
which the Council issued in 1549, when disorders were at their height,
forbidding constables, bailiffs, and head-boroughs to call meetings
except for the purposes required by the law. Hales,[585] as he rode
through the South and Midlands in 1548, was struck by the patience
with which people waited for the Government to take action, and
attributed the disturbances of the ensuing year to the despair caused
by the victory of the local landlords over the Commission, and to the
rejection by Parliament of the Bills which he had introduced. Even
Ket’s campaign in Norfolk, which ended in a sanguinary battle, during
the greater part of it was carried on with an orderliness from which
the Government which suppressed it might profitably have taken a
lesson. Nothing could have been more unlike the popular idea of a
jacquerie. The peasants enjoyed the enormous joke of making the
gentry look foolish a great deal more than cutting their throats, as
during the four weeks in which they were “playing” they might have
done without any difficulty.
“Mr. Pratt, your sheep are very fat,
And we thank you for that;
We have left you the skins to pay your wife’s pins,
And you must thank us for that.”[586]
These lines, pinned on the carcasses of an enclosing landlord’s flocks
and herds, are a fair specimen of their humour. Men may well be
merry together, when they have seen hovering over the fields of an
English county, though but in a fleeting glimpse, the New Jerusalem
where the humble are exalted and the mighty put down; and there is
no inconsistency between such mundane gaiety and the long pent up
passion which on the lips of a nameless labourer burst into the cry,
“As sheepe or lambs are a prey to the wolfe or lion, so are the poor
men to the rich men.”[587] There was much lecturing (the matter is
easily imagined) at the Oak of Reformation, and not on one side only,
for the peasants were tolerant compared with their betters, and a
future archbishop was allowed to address the insurgents on the evils
of their ways; much laying down of hedges and enclosures; much
slaughtering of that beast of iniquity, the man-devouring sheep. There
was none of the massacring of unarmed men which both Henry VIII.
and Elizabeth ordered without compunction when they thought the
times required it, very little of the “making the public good a pretext
for private revenge,” against which the insurgents were warned by
Parker. Though for months after the final tragedy the badges of the
justly-hated Warwick “were not so fast set up but that they were as
fast pulled down" from the city walls, the rebels even in the heat of
their early triumphs claimed only to be executing the Protector’s
Proclamations, and, while indignantly repudiating the name of
traitors, showed a complete readiness to negotiate peaceably with the
Government. The whole movement was less a rising against the State
than a practical illustration of the peasants' ideals, a mixture of May-
day demonstration and successful strike embodied in one gigantic
festival of rural good fellowship. Its bloody termination was, as far as
can be judged, the result of two errors of judgment, one, a
pardonable one, on the part of Ket, the other, unpardonable, on the
part of a nameless member of the other party.[588] When all was over,
and each man reflected after his kind on the great days of Mousehold
Heath, what the camp followers, who attach themselves to every
popular movement, remembered was that for about a month they
had filled their bellies at other people’s expense. "'Twas a merry world
when we were yonder, eating of mutton.” But there were some who,
as they saw Ket swinging on the gallows before the City gates, were
seized with the tumult of pity and hoarse indignation which serves
Englishmen, who are not good at revolutions, in place of the
revolutionary spirit. “O Kette,” one countryman was heard to say to
another, “God have mercy upon thy soul; and I trust in God that the
King’s Majesty and his Councell shall be enformed once between this
and Midsummer evening, that of their own gentleness thou shalt be
taken down and buried, not hanged up for winter store; and set a
quietness in the realm, and that the ragged staff shall be taken down
of their own gentleness from the gentlemen’s gates in this City, and to
have no more King’s arms but one within the City, under Christ.”[589]
The Council, in its gentleness, thought otherwise. Ket still creaked in
his chains, and in the meantime other gallows were rising for other
rebels in Somerset, and Devon, and Cornwall.
What were the aims which at intervals between 1530 and 1560 set
half the counties of England in a blaze? Let us look at the peasants'
programme more closely. It will help us to see the agrarian problem
from the inside. Reduced to its elements their complaint is a very
simple one, very ancient and yet very modern. It is that what, in
effect, whatever lawyers may say, has been their property, is being
taken from them. To be told that social disorders take place because
an envious proletariat aims at seizing the property of the rich would
seem to them a very strange perversion of the truth. They want only
to have what they have always had. They are conservatives, not
radicals or levellers, and to them it seems that all the trouble arises
because the rich have been stealing the property of the poor. Here is
part of a colloquy[590] between Jack of the North beyond the Style,
Robin and Harry Clowte, Tom of Trumpington, Peter Potter, Pyrce
Plowman, and divers other worthies. As will be seen from the verses,
they are birds of night—

Jack. Now for that Slaunder’s sake,

Companye by night I take,
And, with all that I may make,
Cast hedge and ditch in the lake,
Fyxed with many a stake
Though it was never so faste
Yet asondre it is wraste.
* * * * * * *
Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about testbank and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!

ebooksecure.com