The L-Space Web: Analysis

Statistics



Author's note

This work was conducted by someone occasionally called Robert during the last few years in order to satisfy the fans' desire for information regarding deeply moving developments within the DW series, concerning topics like length of the books, and numbers paragraphs and footnotes. You should not ask annoying questions like "why would anybody sit down and count all these words" but just appreciate this hitherto hidden statistical view on Terry's work.

You are free to use the enclosed information for non-commercial purposes, as long as you give this L-space net-page as reference. In scientific works I'd like to be properly cited (vanity, I assume and ask for your pardon). If you wish to use this information for commercial purposes you need to ask permission first.

If you want to contact me, you can do so via email: Robert Neumann.

By The Way, many thanks to Nathaniel.

Enjoy!

Disclaimer

As far as I know, Terry Pratchett has not published any book, apart from "Theatre of Cruelty", in electronic form. This work of statistics does not mean in any way that I possess illegal electronic copies of the DW series or other writing of Pterry. Hence, don't ask me for a copy.

General remarks

If you don't like all this reading, just look at the pictures. They are pretty self-explaining.

This work covers some statistical aspects on Pterry's DW books up to and including TOT and TAMAHER but excluding TLH.

Accuracy: I did my very best to get the correct numbers. However, you cannot rely on this information (that is, you are free to recount :-) to be more accurate than 1 in 1000 for the greater numbers, 1 in 100 for smaller ones and even small numbers may be off by one or two.

What was counted: The numbers are valid for the whole book except the title but including subtitle (as in TCOM, Pyramids or SM). Author's notes and dedications are not considered part of a book (count them for yourself), but the intro in L&L is. Everything else should be (i.e. is considered by me to be :-) fairly obvious.

Graphics: The short stories and novels were counted but not represented in the graphics, because they distort the picture and make it harder to see any trends. They are not really part of any development in length etc. anyway. Not represented are: the short stories Troll Bridge, Theatre of Cruelty and The Sea and Little Fishes, and the short novels (novelettes?) Eric and The Amazing Maurice and His Educated Rodents.

The graphics use normalised numbers, i.e. the actual number was divided by the biggest (or in some cases the smallest) number, resulting in numbers between 0 and 1. If you prefer, multiply them with 100 to obtain percentages. By doing this I was able to fit comparable items, like length in terms of number of characters and length in terms of number of words, in the same diagram although the absolute numbers are quite different.

Finally: Even though I restrained myself and did not add too much interpretation, never forget that you should only very hesitantly trust a statistic that you didn't manipulate yourself!


Normalised length of the books

Due to all those different editions (hardcover, old Corgi, new Corgi, American and Australian ones etc.) it is not possible to compare numbers of pages or lines (I can't afford to buy 'em all in the new Corgi format just to count lines :-). Instead the (hopefully) constant numbers of words and characters [*] were used.

Figure 1

Figure 1

Graph of the Normalised length of the books

The numbers not counting blanks are :
(highest ones high-lightened)

1

TCOM

TLF

TB

ER

Mort

Sour

WS

Pyr

G!G!

TOC

2

307274

297129

14286

300400

331785

364272

387908

399963

453879

5484

3

0.63

0.61

0.03

0.62

0.68

0.75

0.80

0.82

0.93

0.01

4

66014

64371

3242

66260

73018

79222

85583

87931

98530

1243

5

0.61

0.60

0.03

0.62

0.68

0.74

0.80

0.82

0.92

0.01

 

1

Eric

MP

RM

WA

SG

LL

MAA

SM

IT

Mask

2

160925

445217

360560

369779

411903

402190

431711

431580

428415

387411

3

0.33

0.92

0.74

0.76

0.85

0.83

0.89

0.89

0.88

0.80

4

34924

97288

78933

81614

91543

89017

94726

96139

93473

85391

5

0.33

0.91

0.73

0.76

0.85

0.83

0.88

0.90

0.87

0.80

 

1

TSALF

FOC

Hogf

Jingo

CJ

TLC

TFE

TT

TOT

TAMA

2

61238

424453

428310

476792

441270

433707

486097

473408

460096

283276

3

0.13

0.87

0.88

0.98

0.91

0.89

1.00

0.97

0.95

0.58

4

13514

94164

95180

105197

98407

94902

107406

105088

102477

63779

5

0.13

0.88

0.89

0.98

0.92

0.88

1.00

0.98

0.95

0.59

  1. Book (usual abbreviation).
  2. Number of characters in this book (not counting spaces, of course).
  3. Normalised number of characters (multiply by 100 to obtain percentages).
  4. Number of words in this book.
  5. Normalised number of words.

Surprisingly (to me at least) there is no more or less steady climb in the length. Rather, the 7ath novel, G!G! is among the longest, almost followed by one of the shortest (the 10th, RM, is as short as the 5th, Sourcery), and not surpassed before book no. 24, TFE!).

The two lines should match exactly. That they don't is caused by the average length of word varying between 4.48 (CJ) and 4.65 (TCOM) characters per word, a difference of almost 4%. The short stories TB and TOC show far lower values still!)! Obviously, Terry's language changes from book to book. Someone looking for a thesis could give it a look...

Figure 2

Graph of the Average length of words (measured in characters)

Anyway, the numbers upon which fig. 2 is based are:

1

TCOM

TLF

TB

ER

Mort

Sour

WS

Pyr

G!G!

TOC

2

4.65

4.62

4.41

4.53

4.54

4.60

4.53

4.55

4.61

4.41

3

1.00

0.99

0.95

0.97

0.98

0.99

0.97

0.98

0.99

0.95

 

1

Eric

MP

RM

WA

SG

LL

MAA

SM

IT

Mask

2

4.61

4.58

4.57

4.53

4.50

4.52

4.56

4.49

4.58

4.54

3

0.99

0.98

0.98

0.97

0.97

0.97

0.98

0.96

0.98

0.97

 

1

TSALF

FOC

Hogf

Jingo

CJ

TLC

TFE

TT

TOT

TAMA

2

4.53

4.51

4.50

4.53

4.48

4.57

4.53

4.50

4.49

4.44

3

0.97

0.97

0.97

0.97

0.96

0.98

0.97

0.97

0.96

0.95

  1. Book (usual abbreviation).
  2. Average length of the words in characters.
  3. Normalised values of row 2 (multiply by 100 to obtain percentages).

[*] This will present a problem should I ever do Good Omens, as there actually is a difference in length between the English and the American edition (see one of the last annotations for that book). As I don't know of other changes for the American market, nothing else is taken into account.


Normalised numbers of paragraphs and sub-chapters

A sub-chapter is the text body between two empty lines (sometimes filled with one or three asterisks "*"). When you look at TAMAHER (or the non-DW books), where there are true chapters that are still segmented further, the labelling becomes more obvious.

Admittedly, the number of paragraphs belongs to the length of the books (fig. 1) rather than here, but if I include it in fig. 1 it gets all messed up while fig. 3 is rather empty.

Figure 3

Graph of the Normalised numbers of paragraphs and sub-chapters

Numbers are:

1

TCOM

TLF

TB

ER

Mort

Sour

WS

Pyr

G!G!

TOC

2

79

77

4

60

56

73

95

112

115

4

3

0.41

0.39

0.02

0.31

0.29

0.37

0.49

0.57

0.59

0.02

4

2558

3185

190

2884

3538

3894

4198

4042

4732

64

5

0.42

0.53

0.03

0.48

0.59

0.64

0.69

0.67

0.78

0.01

 

1

Eric

MP

RM

WA

SG

LL

MAA

SM

IT

Mask

2

58

150

125

144

137

122

146

196

110

117

3

0.30

0.77

0.64

0.74

0.70

0.63

0.75

1.00

0.56

0.60

4

1738

5372

4982

4610

5377

5331

5866

6044

5291

4964

5

0.29

0.89

0.82

0.76

0.89

0.88

0.97

1.00

0.88

0.82

 

1

TSALF

FOC

Hogf

Jingo

CJ

TLC

TFE

TT

TOT

TAMA

2

14

111

111

107

95

75

89

82

102

49

3

0.07

0.57

0.57

0.55

0.49

0.38

0.46

0.42

0.52

0.25

4

715

4996

5400

5922

5144

4773

5519

5217

5088

2885

5

0.12

0.83

0.89

0.98

0.85

0.79

0.91

0.86

0.84

0.48

  1. Book (usual abbreviation).
  2. Number of subchapters in this book.
  3. Normalised number of subchapters (multiply by 100 to obtain percentages).
  4. Number of paragraphs in this book.
  5. Normalised number of paragraphs.

The number of paragraphs rises steadily till MP (no. 10) and then more or less stays constant.

The number of sub-chapters on the other hand reaches an all-time high (by more than 20 %) at the 16th novel (SM) and than quickly falls down again.


Relative length of the paragraphs and sub-chapters

As we know there are no chapters in the DW books, except for TCOM and Pyramids, and lately TAMAHER. Not even I am willing to do statistics on a sample of three, so I regarded the chapters as sub-chapters for the purpose of fig. 2.

The number of paragraphs and sub-chapters can be expected to rise along with the length of the book. So here we see these numbers divided by the number of characters, thus giving the relative length of the paragraphs and sub-chapters on average [*]:

Figure 4

Graph of the Relative numbers of paragraphs and sub-chapters

We see that the length of paragraphs quite continually falls till SM and then slowly but steadily rises again.

The relative length of the sub-chapters varies far more with a maximum in Mort but almost reaching that level again in the latest novels.

Numbers are:

1

TCOM

TLF

TB

ER

Mort

Sour

WS

Pyr

G!G!

TOC

2

3939

3910

4762

5092

6032

5059

4127

3603

3981

1828

3

0.65

0.65

0.79

0.84

1.00

0.84

0.68

0.60

0.66

0.30

4

120

93

75

104

94

94

92

99

96

86

5

1.00

0.78

0.63

0.87

0.78

0.78

0.77

0.82

0.80

0.71

 

1

Eric

MP

RM

WA

SG

LL

MAA

SM

IT

Mask

2

2823

2988

2908

2586

3029

3324

2977

2213

3930

3340

3

0.47

0.50

0.48

0.43

0.50

0.55

0.49

0.37

0.65

0.55

4

93

83

72

80

77

75

74

71

81

78

5

0.77

0.69

0.60

0.67

0.64

0.63

0.61

0.59

0.67

0.65

 

1

TSALF

FOC

Hogf

Jingo

CJ

TLC

TFE

TT

TOT

TAMA

2

4711

3859

3894

4498

4694

5861

5524

5845

4555

5902

3

0.78

0.64

0.65

0.75

0.78

0.97

0.92

0.97

0.76

0.98

4

86

85

79

81

86

91

88

91

90

98

5

0.71

0.71

0.66

0.67

0.71

0.76

0.73

0.76

0.75

0.82

  1. Book (usual abbreviation).
  2. Number of characters per subchapter in this book (i.e. length of subchapter measured in characters).
  3. Normalised length of subchapters (multiply by 100 to obtain percentages).
  4. Number of characters per paragraphs in this book.
  5. Normalised length of paragraphs.

[*] Technically you could divide the number of paragraphs and sub-chapters by the number of words, too, because, as we have seen in fig 1, it is not exactly proportional to the number of characters. However, the difference is too small as to make any difference, so to speak.


Footnotes

Everybody likes them. The one-liner, the two-liner, the longer explanations... Always good for a laugh or a smile, a chuckle or even what the ardent fans call a ROTFLKAS [*]. I love them.

Reason enough to look for a development in the footnotes. Are they getting more, or less, or staying constant? I put some effort into researching this on a statistical basis.

To cut it short: I can show you nothing. For whatever I looked, the picture is always the same: No development, no trend, no correlation. It is presented below, since I had done it. But the worth is dubious, you may as well skip this section. Quite depressing for one who set out to discover hidden wonders in the world of numbers. Anytime soon now I will go looking for a real life (in addition to the one I am leading already :-).

[*] Roll on the floor laughing, kicking and screaming.

Number of the footnotes

The first attempt to detect an ordering development in the footnotes was to research how many of them there are. But the number of footnotes varies quite erratically. Little is gained if we take the length of the books into account (i.e. divide the number of footnotes by the length of books measured in number of characters [*]):

Figure 5

Graph of the Normalised numbers of footnotes

Numbers are:

1

TCOM

TLF

TB

ER

Mort

Sour

WS

Pyr

G!G!

TOC

2

1

5

1

1

12

25

24

33

24

1

3

0.02

0.11

0.02

0.02

0.27

0.57

0.55

0.75

0.55

0.02

4

1.51

7.77

30.85

1.51

16.43

31.56

28.04

37.53

24.36

80.45

5

0.03

0.16

0.62

0.03

0.33

0.64

0.57

0.76

0.49

1.63

 

1

Eric

MP

RM

WA

SG

LL

MAA

SM

IT

Mask

2

13

29

19

24

10

44

24

30

25

9

3

0.30

0.66

0.43

0.55

0.23

1.00

0.55

0.68

0.57

0.20

4

37.22

29.81

24.07

29.41

10.92

49.43

25.34

31.20

26.75

10.54

5

0.75

0.60

0.49

0.59

0.22

1.00

0.51

0.63

0.54

0.21

 

1

TSALF

FOC

Hogf

Jingo

CJ

TLC

TFE

TT

TOT

TAMA

2

0

16

26

18

14

24

21

14

15

4

3

0.00

0.36

0.59

0.41

0.32

0.55

0.48

0.32

0.34

0.09

4

0.00

16.99

27.32

17.11

14.23

25.29

19.55

13.32

14.64

6.27

5

0.00

0.34

0.55

0.35

0.29

0.51

0.40

0.27

0.30

0.13

  1. Book (usual abbreviation).
  2. Number of footnotes.
  3. Normalised number of footnotes (multiply by 100 to obtain percentages).
  4. Number of footnotes per 10.000 words (per 10.000 because otherwise the numbers would be annoyingly small) in this book.
  5. Normalised of footnotes per 10.000 words.

The number of footnotes per word closely matches the number of footnotes per book, so this again provides no clue for any change in the master's mind. (May I suggest a thesis looking into possible differences in the number of footnotes in the "lighter" and more serious novels?)

A statistical trapdoor is TOC: it contains one footnote, but because it is so short it has the highest number of footnotes per word by far (for the normalisation he value for L&L was used).

Row 5 of the tables can be translated into "How many words there are in-between two footnotes", for your interest this is given below (obtained by dividing the number of words by the number of footnotes):

1

TCOM

TLF

TB

ER

Mort

Sour

WS

Pyr

G!G!

TOC

2

66014

12874

-

66260

6085

7922

3423

2665

4479

1243

 

1

Eric

MP

RM

WA

SG

LL

MAA

SM

IT

Mask

2

2910

3355

3947

3265

9154

2023

3266

3205

3595

9488

 

1

TSALF

FOC

Hogf

Jingo

CJ

TLC

TFE

TT

TOT

TAMA

2

-

5885

3661

5844

3954

7029

4670

7506

7320

15945

  1. Book (usual abbreviation).
  2. Average number of word in-between two footnotes.

[*]See footnote for Relative length of the paragraphs and sub-chapters[*]

Total length of footnotes

OK, I thought, let's see if the picture changes, if I consider the total number of words and characters in the footnotes. Maybe changes in the length balance for changes in numbers... But, alas, nothing balances, see for yourself:

Figure 6

Total Length of footnotes

Numbers are:

1

TCOM

TLF

TB

ER

Mort

Sour

WS

Pyr

G!G!

TOC

2

364

223

0

151

801

466

964

1345

688

18

3

0.22

0.13

0.00

0.09

0.48

0.28

0.58

0.81

0.41

0.01

4

1677

1140

0

734

3714

2186

4419

6342

3202

71

5

0.22

0.15

0.00

0.10

0.50

0.29

0.59

0.85

0.43

0.01

 

1

Eric

MP

RM

WA

SG

LL

MAA

SM

IT

Mask

2

663

777

1128

925

207

1348

1165

1087

726

330

3

0.40

0.47

0.68

0.55

0.12

0.81

0.70

0.65

0.44

0.20

4

3189

3580

5546

4312

898

6214

5551

4940

3377

1575

5

0.43

0.48

0.74

0.58

0.12

0.83

0.74

0.66

0.45

0.21

 

1

TSALF

FOC

Hogf

Jingo

CJ

TLC

TFE

TT

TOT

TAMA

2

0

615

1667

684

1312

770

547

577

657

155

3

0.00

0.37

1.00

0.41

0.79

0.46

0.33

0.35

0.39

0.09

4

0

2939

7492

3234

6023

3555

2525

2733

2864

641

5

0.00

0.39

1.00

0.43

0.80

0.47

0.34

0.36

0.38

0.09

  1. Book (usual abbreviation).
  2. Number of words in the footnotes.
  3. Normalised number of words in the footnotes (multiply by 100 to obtain percentages).
  4. Number of characters in the footnotes.
  5. Normalised number of characters in the footnotes.

Length of words in footnotes

The difference in the length of footnotes measured in words and characters is again [*] due to the differences in the length of the words. The differences are greater than for the total text (that is probably due to the smaller sample, but offers, of course, a wide field of interpretation :-), and generally, but not always, the word used in footnotes are longer then the ones used in the total text.

Figure 7

Graph of the length of words in footnotes

Numbers are:

1

TCOM

TLF

TB

ER

Mort

Sour

WS

Pyr

G!G!

TOC

2

4.65

4.62

4.41

4.53

4.54

4.60

4.53

4.55

4.61

4.41

3

4.61

5.11

-

4.86

4.64

4.69

4.58

4.72

4.65

3.94

4

0.91

0.90

0.86

0.89

0.89

0.90

0.89

0.89

0.90

0.86

5

0.90

1.00

-

0.95

0.91

0.92

0.90

0.92

0.91

0.77

 

1

Eric

MP

RM

WA

SG

LL

MAA

SM

IT

Mask

2

4.61

4.58

4.57

4.53

4.50

4.52

4.56

4.49

4.58

4.54

3

4.81

4.61

4.92

4.66

4.34

4.61

4.76

4.54

4.65

4.77

4

0.90

0.90

0.89

0.89

0.88

0.88

0.89

0.88

0.90

0.89

5

0.94

0.90

0.96

0.91

0.85

0.90

0.93

0.89

0.91

0.93

 

1

TSALF

FOC

Hogf

Jingo

CJ

TLC

TFE

TT

TOT

TAMA

2

4.53

4.51

4.50

4.53

4.57

4.48

4.53

4.50

4.49

4.44

3

-

4.78

4.49

4.73

4.59

4.62

4.62

4.74

4.36

4.14

4

0.89

0.88

0.88

0.89

0.89

0.88

0.89

0.88

0.88

0.87

5

-

0.93

0.88

0.92

0.90

0.90

0.90

0.93

0.85

0.81

  1. Book (usual abbreviation).
  2. Length of words in the total text (including footnotes), measured in characters.
  3. Length of words in the footnotes, measured in characters.
  4. Normalised length of words in the total text (multiply by 100 to obtain percentages).
  5. Normalised length of words in the footnotes.

[*] Cf. fig 2 and it's description

Average length of individual footnotes

While we are about it, let's see if there is a development in the length of the footnotes, even if it does not cancel out the development in their number. Let's see, if they tend to get one-liners, or long explanations of some Discworld detail or other.

Well, it is not the case. Firstly I had to leave out TCOM and ER for the graphics, because their only one but long footnote (you can see it in the table: Their footnote is 5.5 respective 2.4 times longer then the next-highest average) just distort the picture. Secondly, the highest averages (Mort, RM and Hogfather) are about three times higher than the shortest (SG and TFE). Thirdly, not a trace of a trend is recognisable.

Figure 8

Graph of the Average length of individual footnotes

Numbers are:

1

TCOM

TLF

TB

ER

Mort

Sour

WS

Pyr

G!G!

TOC

2

364

45

0

151

67

47

39

41

31

18

3

5.43

0.67

0.00

2.25

1.00

0.70

0.58

0.61

0.47

0.27

4

1677

228

0

734

310

219

177

192

146

71

5

5.41

0.74

0.00

2.37

1.00

0.71

0.57

0.62

0.47

0.23

 

1

Eric

MP

RM

WA

SG

LL

MAA

SM

IT

Mask

2

55

27

56

37

21

31

40

36

28

37

3

0.82

0.40

0.84

0.55

0.31

0.46

0.60

0.54

0.42

0.55

4

266

123

277

172

90

141

191

165

130

175

5

0.86

0.40

0.89

0.56

0.29

0.46

0.62

0.53

0.42

0.56

 

1

TSALF

FOC

Hogf

Jingo

CJ

TLC

TFE

TT

TOT

TAMA

2

0

38

64

38

55

55

24

41

47

39

3

0.00

0.57

0.96

0.57

0.82

0.82

0.35

0.62

0.70

0.58

4

0

184

288

180

251

254

110

195

205

160

5

0.00

0.59

0.93

0.58

0.81

0.82

0.35

0.63

0.66

0.52

  1. Book (usual abbreviation).
  2. Average length of footnotes, measured in words.
  3. Normalised average length of footnotes, measured in words (multiply by 100 to obtain percentages).
  4. Average length of footnotes, measured in characters.
  5. Normalised average length of footnotes, measured in characters.

Length of footnotes versus length of books

My last attempt to obtain insight in the footnote-topic was: to see if there is something less erratically to observe if I divide the number of words/characters in the total text by the number of words/characters in the footnotes only. I.e. how many times is the book (total text) longer than the footnotes.

Here at last there is at least something to be observed. The highest values by far show ER and SG, followed (if separated by a factor of two) by Maskerade. On the other end of the value-spectrum, there are ten books showing values between 0.1 and 0.2.

I leave the interpretation to you, just bear in mind that a high value equals a high text-to-footnote-ratio, i.e. few footnotes.

Figure 9

Graph of the Length of footnotes versus length of books

Numbers are:

1

TCOM

TLF

TB

ER

Mort

Sour

WS

Pyr

G!G!

TOC

2

181

289

0

439

91

170

89

65

143

69

3

0.41

0.65

0.00

0.99

0.21

0.38

0.20

0.15

0.32

0.16

4

183

261

0

409

89

167

88

63

142

77

5

0.40

0.57

0.00

0.89

0.19

0.36

0.19

0.14

0.31

0.17

 

1

Eric

MP

RM

WA

SG

LL

MAA

SM

IT

Mask

2

53

125

70

88

442

66

81

88

129

259

3

0.12

0.28

0.16

0.20

1.00

0.15

0.18

0.20

0.29

0.59

4

50

124

65

86

459

65

78

87

127

246

5

0.11

0.27

0.14

0.19

1.00

0.14

0.17

0.19

0.28

0.54

 

1

TSALF

FOC

Hogf

Jingo

CJ

TLC

TFE

TT

TOT

TAMA

2

0

153

57

154

72

128

196

182

156

411

3

0.00

0.35

0.13

0.35

0.16

0.29

0.44

0.41

0.35

0.93

4

0

144

57

147

72

124

193

173

161

442

5

0.00

0.31

0.12

0.32

0.16

0.27

0.42

0.38

0.35

0.96

  1. Book (usual abbreviation).
  2. Number of words in the total text (including footnotes) divided by number of words in the footnotes (i.e. "How many times is the book longer than the footnotes").
  3. Values of row 2, normalised (multiply by 100 to obtain percentages).
  4. Number of characters in the total text (including footnotes) divided by number of words in the footnotes
  5. Values of row 4, normalised.

[Up]
This section of L-Space is maintained by esmi

The L-Space Web is a creation of The L-Space Librarians
This mirror site is maintained by A.H.Davis