Transcription and Markup Conventions

The WSC's transcription system is orthography based. The conventions used in the release version of the WSC, along with the markup conventions used are outlined in this section. All markup is sgml conformant and is based on the system used by ICE (Nelson 1993). A summary of the conventions is provided in Appendix 3.

1. Character Set

Alphabetic roman characters are used in lexical transcription and editorial comments. No diacritics or non-roman characters are permitted. No punctuation is used (except for apostrophes) and upper case is reserved for marking emphatic stress. All markup is sgml conformant and appears between angle brackets < >. Markup symbols utilise a wider range of characters than transcription. The permissible characters used in the marked corpus are:

a b c d e f g h i j k l m n o p q r s t u v w x y z

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

? [ { . , ,, & # : 0 1 2 3 4 5 6 7 8 9

2. Referencing system

The beginning and end of each file are marked by <I> and </I> respectively. After this each text/intonation unit is marked by a reference e.g., <WSC#DPC334:0005:LR>. These references include the following information:

WSC

The first three characters in every reference indicate that the file is from the WSC

#

Phrase marker

DPC334

Extract code

0005

Phrase/unit number

LR

Speaker Identification

This system gives a unique identifier to every phrase. It is designed as a referencing system. The numbering was originally done in fives, leaving the in-between numbers to allow transcripts to be modified without having to renumber all the phrases.

References are also provided for speech from ineligible speakers, although this is marked as extra corpus text (see section 14.8, Extra Corpus Text).

From these references, users of the corpus can not only specify exactly which phrases they are discussing, but they can easily retrieve all available information on that phrase (e.g., speaker details).

To make the explanations of other notation clearer in the rest of this section, references are not generally added to the examples. Section 14.10, Transcript Examples, provides some examples where references have been added.

2.1. Speaker Identification

The speaker identification section of each reference is always two characters. In the case of broadcast material, speakers' real initials are generally used. In non-broadcast categories, the initials of speakers' pseudonyms are used (see section 14.3, Pseudonyms).

There are exceptions to this. In some categories, the first character in the speaker code is used to denote the kind of relationship between the participants in the interaction. For example, in the broadcast interview category, the interviewer always has I as the first character in their speaker code, e.g., IJ or IP. Information on particular characters being used in this way is provided at the beginning of each category section in section 15, Texts.

Another exception is when speakers are ineligible. Ineligible speakers always have X as the first initial, e.g., XX or X1.

The final exception to this is when speakers are eligible, but we do not have detailed demographic information on them. This applies mainly to the DGB Radio talkback category (see section11.1.1, Who counts as a New Zealander?). These speakers always have Z as the first character in their speaker code, e.g., ZZ or Z1.

As these speaker identification characters are not unique to speakers, each individual speaker has been assigned a unique speaker code. These codes are numeric and are listed in section 15, Texts, along with the speaker identification characters. All participants are also listed by their unique speaker code in section 16, Participants.

3. Pseudonyms

Pseudonyms are used to label speakers and people talked about, unless their name is a matter of public record. Generally, a name with the same gender or ambiguity of gender, stress patterns, number of syllables and ethnicity is used.

E.g.,

REAL NAME

PseUdonym

Sue

Jill

Chris(tina)

Pat(ricia)

Steve(n)

Rich(ard)

Tama

Hemi

4. Discourse Features

4.1. Stress

CRAZY, UNbelievable Capitals are used to indicate emphatic stress.

4.2. Question intonation, where unclear

? Used to signal an interrogative where it is ambiguous on paper.

E.g., "you're going to the zoo tomorrow?" (question) versus "you're going to the zoo tomorrow" (statement).

Where the sentence structure is unambiguous, no punctuation is added, e.g. are you going to the zoo tomorrow

4.3. Comprehension Problems and Transcriber Doubt

<?>well</?> Transcriber's best guess at unclear speech

<unclear> </unclear> Untranscribable or incomprehensible speech

When there is doubt about the accuracy of transcription the markup symbols <?> </?> are used. When the speech is untranscribable or incomprehensible the markup symbols <unclear> </unclear> are used. In the later case, the ICE system requires a guess be made at the number of words involved. WSC transcripts generally just say <unclear>word</unclear>, as when speech is unclear it is hard to specify the number of words involved.

4.4. Incomplete words

<.>wha</.> Incomplete word

The symbols <.> </.> indicate that a word has been cut off, both self-interruption and other speaker interruption.

4.5. Pauses

<,> pause of up to or equal to one second

<,,> pause over one second and up to two seconds

<,,><&>3</&> pause of over two seconds and up to three seconds

The ICE system only notes whether a pause is short or long. In the WSC we use their short pause symbol, <,>, to denote a short pause of up to or equal to one second and their long pause symbol, <,,>, to denote a pause over one second and up to two seconds. For pauses longer than this the long pause symbol is used <,,> along with the number of seconds surrounded by the editorial comment markup symbols <&> </&> (see section 14.5, Editorial Comments, for further uses of these symbols). For example <,,><&>3</&> denotes a pause of over two seconds and up to three seconds and <,,><&>4</&> denotes a pause of over three seconds and up to four seconds.

4.6. Latching

<latch> Indicates latching, i.e., where the second utterance immediately follows the first with no discernible pause or overlap.

E.g., A: do you know what the time is<latch>

B: six o'clock

4.7. Simultaneous speech

<{> <[> Indicates start of simultaneous or overlapping speech in utterance of designated current or first speaker.

</[> Indicates end of simultaneous or overlapping speech in utterance of current or first speaker.

<[> Indicates start of simultaneous or overlapping speech in utterance of incoming or second speaker.

</[> </{> Indicates end of simultaneous or overlapping speech in utterance of incoming or second speaker.

E.g., A: i'd like to come as well <{><[>is</[> that okay

B: <[>yeah</[></{>

Our original transcription system had been syllable sensitive, i.e., overlaps were sometimes marked in the middle of a word,

e.g., A: dis<{><[>gusting</[>

B: <[>yeah</[></{>

The ICE system of markup, along with many software programmes for analysing corpora, requires word integrity. For the release version of the corpus, therefore, all overlaps have been moved to ensure word integrity,

e.g., A: <{><[>disgusting</[>

B: <[>yeah</[></{>

4.8. Multiple intra-turn Overlaps

Numbering is added where a speaker is overlapped more than once within a turn:

Eg.

A: you deep fried them <{1><[1>did</[1> you or just <{2><[2>pan fried them</[2>

B: <[1>mm</[1></{1>

B: <[2>no<,> they were</[2></{2> very greasy

Sometimes two speakers may overlap a third at the same time. This has been marked as follows, with <{> and </{> encompassing the whole overlap:

e.g. A: it all turned out <{><[>all right</[> in the end

B: <[>yeah</[>

C: <[>mm</[></{>

4.9. Numbering of overlaps: consecutive and disjoint numbering

A: remember that time we went to gisborne <{1><[1>with</[1> the martins and we went to that <{2><[2>place with the huge</[2> rocks

B: <[1>yeah</[1></{1>

C: <[2>and it rained</[2></{2>

C: and it <{1><[1>rained</[1> the whole time <{2><[2>and none of us</[2> had raincoats

A: <[1>yeah</[1></{1>

A: <[2>i'd forgotten that</[2></{2>

Although the two interruptions on A are made by different people (first B, then C), they are still numbered consecutively as interruptions made within a single turn of A's.

When the speaker changes (in this case from A to C), and a different person is being interrupted (in this case C), the numbering of interruptions on C starts from 1 again.

4.10. Paralinguistic anthropophonics

<O> </O> Paralinguistic anthropophonics

Laughter, coughing, sniffing, snorting, sneezing and other such noises are marked <O> </O>.

E.g., A: <O>groans</O> <O>laughs</O> <O>clears throat</O>

In all the monologue categories coughing, laughter etc from the audience is not transcribed unless it obliterates or interrupts the speaker.

The most frequent noises marked with the symbols <O> </O> are:

<O>tut</O> bilabial/alveolar/dental clicks

<O>voc</O> other vocalisations not covered by any other convention.

4.11. Features occurring through sections of speech

Features which apply to sections of speech are generally noted in angle brackets around the speech they apply to. E.g. "<quickly>we won't talk about that</quickly>", "<drawls>um</drawls> but i think"

Annotations use capitals where appropriate (proper nouns, etc) and follow standard spelling. Comments take the form of adverb, third person singular, or "in/with a ____". E.g.,

A: <quietly>yes</quietly> <whispers>i know</whispers> <with pseudo American accent>have a nice day</with pseudo American accent>.

Some of the most common features marked in this way are:

<drawls> </drawls> <exhales> </exhales>

<inhales> </inhales> <laughs> </laughs>

<paraphrases> </paraphrases> <quickly> </quickly>

<quietly> </quietly> <reads> </reads>

<shouts> </shouts> <sighs> </sighs>

<sings> </sings> <whispers> </whispers>

<with pseudo language accent> </with pseudo language accent>

<with silly voice> </with silly voice>

Note that <laughs> </laughs> is used when laughter occurs over an utterance (to indicate voice quality where laughter and speech are co-extensive),

e.g., "<laughs>oh i can be like that too</laughs>"

As seen in section 14.4.10, Paralinguistic Anthropophonics, normal laughter, i.e., when it does not occur over an uttereance, is marked <O>laughs</O>.

Note that as with overlap markup, markup of this type encompasses the whole word, even if the tag does not apply to that whole word (e.g. "<laughs>yesterday</laughs>" and NOT "<laughs>yester</laughs>day").

5. Editorial Comments

Markup of editorial comments involves the use of the symbols <&> </&>. These symbols are used in a range of situations, e.g., as seen earlier to mark pauses. Some of the situations they are used in are outlined below.

5.1. Background Conversation and Noise

Where two conversations proceed simultaneously and transcribers were able to transcribe one as the "main" conversation, only that conversation has been transcribed and a note added for the background conversation: e.g. <&>conversation between AP and MM continues in background - not transcribed</&>

In all the monologue categories questions from the audience (in lectures, demonstrations, etc.) may be transcribed but are not included in the word count. Coughing, laughter etc. from the audience is not transcribed unless it obliterates or interrupts the speaker (see section 14.4.10, Paralinguistic Anthropophonics).

Background noise throughout a transcript is noted in the transcribers' comments in the database tables (see section 17, Database Files). Isolated noises, e.g. cups banging on tables, doorbells, sirens, are only transcribed if the participants acknowledge the sound by speech or action, or where speech is obscured by the noise.

5.2. Other Contextual Information

Other contextual information that may be useful is also noted using the <&> </&> markup. E.g., <&>reads letter to self</&>, <&>goes to get book</&>.

5.3. Times

Where the extract begins on the tape is noted at the beginning of the body of a transcript. The approximate position of every whole minute is also noted. All times are tagged with the editorial comment markup <&> </&>.

E.g. (adapted from MSN062)

 

<&>[side one]</&>

 

<&>6:44</&>

M: and now join me for <indig=Maori>nga korero o te wa</indig=Maori> a

 

round up of maori news events the funeral of a long time battler for maori land claims was held during the week wanganui river <indig=Maori>kaumatua</indig=Maori> hikaia amohia died at <&>7:00</&> his home in taumarunui aged seventy five and was buried at the te peka <indig=Maori>urupa</indig=Maori> near the king country town…

5.4. Non Standard Pronunciations

Speech errors (including unexpected pronunciations, malapropisms, spoonerisms) are transcribed as the orthographic standard, with an editorial comment where appropriate.

e.g. ....talking about trench warfare <&>pronounced as warfore</&>...

e.g. ...and one of my nephews <&>pronounced nebjuz</&>...

Brought pronounced as bought is marked like this:

...brought <&>pronounced as bought</&>...

Widespread New Zealand English non-standard pronunciations are not given a special note.

e.g. titahi (place name) pronounced as /titai/

e.g. merry to rhyme with mary

e.g. women pronounced the same as woman

5.5. Phonetics

Forms of vowels (used in discussions of phonetics) are transcribed as editorial comment using keywords from the KIT set. Use of IPA symbols is not permitted in WSC transcripts (see section 14.1, Character Set).

For example "<&>says GOAT diphthong</&>" or "<&>says DRESS vowel</&>". (The only transcript in the corpus using phonetic mention forms is MST051.) The standard lexical set is taken from Wells (1982:xviii-xix).

Monophthongs

KIT DRESS TRAP LOT

STRUT FOOT BATH CLOTH

NURSE FLEECE PALM THOUGHT

GOOSE START NORTH FORCE

Diphthongs

FACE GOAT PRICE CHOICE

MOUTH NEAR SQUARE CURE

 

6. Utterances in Languages Other Than English

Maori was the most common language other than English used by participants and is discussed separately below. Other languages were also used by contributors to the corpus, though many of the examples from other languages occurred as personal or place names (e.g., Ceau escu, Azerbaijan), or as specialist terminology (e.g., bok choy, fritto misto). Where a transcript contains words or phrases that are not considered part of New Zealand English and are not Maori, these have been marked <foreign> </foreign>.

e.g. <foreign=German>wunderbar</foreign=German>

Non-native, failed or half-hearted attempts to pronounce other languages also created difficulties. Speakers often resort to stereotyped accents for dramatic or humorous effect, including Franglais and mock-American accents. We annotated imitation as "<with pseudo language accent> </with pseudo language accent>" (see section 14.4.11 Features Occurring Through Sections of Speech), while code-switching or lexical borrowing was marked as "<foreign=language> </foreign=language>".

7. Maori

In NZE some speakers code-switch between Maori and English. There are also some words of Maori origin that are part of NZE. Short phrases in Maori are transcribed for the corpus and are marked <indig=Maori> </indig=Maori>, e.g., <indig=Maori>marae</indig=Maori>. Frequently occurring Maori words used in the WSC transcripts are glossed in section 18, Maori Glossary. Low frequency items are glossed in the Extracts database file (see section 17, Database Files).

Proper names including those of flora and fauna are glossed but not annotated. Eg. (i) paua, kauri and tui are unmarked but glossed. The exception to this is when there is an English equivalent and the English equivalent is frequently used in New Zealand English. E.g. kahikatea would be marked "<indig=Maori>kahikatea</indig=Maori>" and glossed because white pine is commonly used to refer to podocarpus excelsum.

Maori tribal names are also not marked as <indig=Maori> </indig=Maori>. For example, "ngati whatua" for Ngaati Whaatua.

Longer sections of uninterrupted Maori are omitted from the transcripts, but are summarised to enable the sense of the discourse to be followed. These summaries are marked with the general editorial comment markup, e.g. <&>Section in Maori where A and B discuss A's father's recent ill health and operation</&>.

Transcribers with a knowledge of Maori were essential for some samples. The bulk of Maori in the corpus, however has wide currency in New Zealand, even in the speech of monolingual speakers of English. All such speakers could be safely assumed to know the words Maori, Pakeha and kiwi, and most would recognise kohanga reo, tino rangatiratanga and aroha (see Deverson 1994 and Bellett's 1995 study of the extent of Maori borrowing into New Zealand English).

7.1. Vowel Length in Maori

Vowel length is phonemic in Maori. The accepted way of transcribing longer vowels in Maori is to use a macron. The macron, however, is not a permissible character in the ICE system. For the release version of the WSC all vowels have been transcribed as single vowels. In section 18, Maori Glossary, macrons have been used.

Williams (1971), A Dictionary of the Maori Language was consulted for standard spelling.

7.2. Plurals, Clitics and Inflectional Endings

This note applies to any language in which English morphemes are added to non-English bases, but is most relevant to Maori. If a clitic (e.g. possessive 's) or an inflectional or plural ending is added to a non-English word (e.g. a couple of tangis) then the whole orthographic word was enclosed within markup. Eg. "a couple of <indig=Maori>tangis</indig=Maori>".

8. Extra Corpus Text

Transcribed words from ineligible speakers are marked with <X> </X>.

E.g.,

AR: and he was okay about it?

<X>

XD: after the fifth drink or so</X>

As noted in section 14.2.1, Speaker Identification, all ineligible speakers can be identified by the use of X as the first character in their speaker identification.

9. Spelling

As noted above, the transcription system of the WSC is an orthographic one, using conventional English spelling to represent naturally occurring speech (as opposed to a phonetic transcription system such as the International Phonetic Alphabet or eye-dialect, which attempts to capture features unique to a dialect or spoken language generally). The decision to use conventional spellings does not, however, solve all problems. A number of spelling options exist within prescribed usage. Some of the options and the decisions made in relation to the WSC are outlined below.

9.1. American versus British spelling

British English has historically had greater influence over New Zealand English than American English. New Zealand spellers generally opt for British spellings rather than American ones. While possible American influence is evident in many areas of spoken New Zealand English (see, for example Vine 1995), New Zealand orthography is still resisting American innovation. Australia is more willing to accept, for example, -or spellings where British spelling would have -our, e.g., in words like colo(u)r and humo(u)r. Peters (1995: 546) reports that in the written Australian Corpus of English, -or spellings outnumber those with -our by a ratio of between 1.4 to one and 9.9 to one. A search of the WWC by Robert Sigley revealed only one clear example of an -or spelling (vaporized) out of 1552 tokens. (A total of nine other tokens with -or spellings were found. These were all discounted because they were either an author's name, a probable spelling error or occurred in texts set or edited in America.) Sigley concluded that overall there was no systematic spelling influence. There were some isolated words where influence was evident, but these were more often in American dominated domains, for example, computing (Sigley 1997).

Selecting a single orthographic form for a lexeme reduces the number of spelling variants one has to search for when trying to locate, for example, relevant phonological environments. For these reasons, the British English preference and search efficiency, spelling variation has been standardised on British English conventions.

9.2. Word Division

Word division is another area which obviously affects word searches and word counts. A no-hyphen policy was established for corpus transcription; the hyphen was reserved for incomplete words in our original transcription system (a very common speech phenomenon). This dramatically reduced the number of decisions over the hyphenation of compound forms. All that remained was a decision of one word form versus two, e.g. byproduct or by product. The general convention adopted was to write these as one word form.

In both theoretical and practical terms the choice of word division for a corpus is not just a trivial one of house style. The sampling size of corpus extracts and the corpus as a whole is measured in number of orthographic words (the one million words of the WSC consists predominantly of 2,000 word extracts). Whether or not the large number of compound forms are transcribed as one or two orthographic words affects the length of the corpus. A lack of standardisation of word division between extracts results in a lack of parity between the length of extracts. Obviously, if extracts differ in the number of words they contain, this should not be solely the result of inconsistent spelling.

9.3. Spelling decisions

Our first reference on spelling was to the Collins English Dictionary (Makins 1994) as it has a good coverage of spoken, contemporary and colloquial English. If a word was not listed there, a variety of other sources was consulted, such as New Zealand dictionaries (Orsman 1989, Orsman and Orsman 1994), the Oxford English Dictionary (Simpson and Weiner 1989), specialised dictionaries of technical terms or primary sources.

On the basis of these sources a standard spelling was determined. The decision was then entered into the transcribers' manual.

Hapax legomena such as matery (mate + ery) or rare inflections such as relativities (relativity + s) were spelt according to conventional (British) spelling rules.

9.4. Non-standard Speech

Below is an alphabetical list of some of the spelling decisions made in relation to the transcription of the WSC. Noises and agreement/disagreement forms are treated separately in sections 14.9.5 and 14.9.6 respectively.

GENERAL PRINCIPAL:

Non-standard speech has been transcribed in standard orthographic form closest to the full morpheme so that it can be picked up in word frequency counts. Exceptions are very frequent variants with familiar (standardised) variant forms, see below, e.g. cos, gonna but he not 'e and stamping not stampin'.

SOURCES: CED = Collins English Dictionary; HNZD = Heinemann NZ Dictionary; NZED = The New Zealand English Dictionary; NZSB= The NZ Style Book; ODMS = Oxford Dictionary of Modern Slang, OEDAS = Oxford English Dictionary Additions Series.

about 'bout for about is considered standard pronunciation and is transcribed about, without annotation

and represents all variants. E.g. and, 'n, 'nd etc

anti irish } anti- compounds are transcribed with space between

anti fur } elements

arised transcribed as "arised <&>past tense of arise</&>"

bikkies contraction of biscuits

bitsy fragmented, in bits

blimmin colloquial use of blooming, as in blimmin heck

blottoed drunk, unconscious from drink (blotto, CED)

bogan white trash

boomph rhymes with oomph

bumph rhymes with month (when said with a final voiceless labiodental fricative) - derogatory reference to unnecessary official forms, etc

byproduct by-product

cos represents all abbreviated variants of because

cuz/cuzzy contraction of cousin to be spelt cuz, cuzzy, or

(bro/s) cuzzy bro(s) where appropriate (NZED)

doozy something of surprising excellence (ODMS)

dork a fool, penis (HNZD, ODMS)

eh tag. E.g. "badjelly is really cute eh".

fella(s) with final schwa represents colloquial use of fellow(s)

flattie(s) contraction of flatmate(s), flat tyre(s) (HNZD)

fundie contraction of fundamentalist (ODMS)

gaw for God pronounced /go/ (ODMS)

gawd for God pronounced /god/ (ODMS)

gonna pronunciation of going to. However all other non-standard verb forms have been transcribed in full (e.g. "hafta" as have to, "wanna" as want to, etc)

goober an idiot (same pronunciation, /gub /, sourced in CED, but meaning "peanut")

Hawke Bay the coast from Mahia Peninsula to Cape Kidnappers (NZSB, p74)

Hawke's Bay the province (NZSB, p74)

hodgepodge (AmE) } Both forms used as required, depending on the

hotchpotch (BrE) } pronunciation (CED)

hoon lout, hooligan (HNZD)

humongous Sourced in ODMS and OEDAS, humongous is preferred spelling as it reflects the word's possible origin as blend of huge(ous) and monstrous

injust transcribed as "injust [ie, unjust]"

jeez for the contracted form of jesus

me in the example "put it in me mouth" is transcribed as me and marked as "me [pronunciation of my]"

-oey -y suffix on words ending in -o take a linking -e-, e.g. tomatoey and potatoey

okay is the standard spelling

okey doke(y) rhyming form of okay

rellies contraction of relatives (NZED)

righty oh based on right oh

should've either should have or should've is used, even when should of is said with a distinct full vowel.

strewth expression of dismay or surprise (CED)

sweetie term of endearment, lolly (CED)

them 'em for them is considered standard pronunciation and is transcribed them, without annotation

thingumabob } for person/thing whose name is unknown,

thingumajig } forgotten, or deliberately overlooked (CED)

thingummy } (NB, not thingamajig or thingamabob)

whatsitsname }

whatshername } for person/thing whose name is unknown,

whatshisname } forgotten, or deliberately overlooked

whatsit }

wus piker, coward (spelling sourced in CED as South Wales

casual term of address)

wussie piker, coward (unsourced)

yay (big) expression of triumphant delight, also as in yay big (this high) (ODMS)

yeehah expression of excitement (unsourced)

yo a greeting (ODMS)

yukky adjectival form of yuk, an expression of disgust (CED)

9.5. Noises

ah represents the vocalisation in expressions like "ah, I see what you mean now" i.e. NOT a hesitation marker. The vowel may be relatively long or short, but is transcribed as ah regardless.

aha yes

er all hesitations not ending in -m

um all hesitations ending in –m

mm minimal feedback

mhm yes

oh represents all utterances in the oh - ooh group

uh huh uh-huh, expression of agreement or acknowledgement (CED)

uh uh no

whoohoo expression of enthusiastic excitement

whoops exclamation of apology or surprise (CED)

9.6. Agreement/Disagreement Forms

yes no }

yeah nah } used to transcribe variants of yes/no.

yep nope }

9.7. Contractions

'll, 've, 'd, n't etc are used as appropriate, regardless of whether the clitic is attached to a verb or noun host. For example "the fellas've done it before". Auxiliary clitics can also appear outside negative clitics, e.g. "she mustn't've been conscious at the time". (See also gonna, should've, blimmin, cuz, rellies, jeez above.)

9.8. Numbers, Acronyms and Abbreviations

DOC (Department of Conservation) can be said as either /dok/ or /di ou si/, transcribed as doc or d o c respectively. RSA is transcribed "r s a" and MPs "m ps", etc.

Numbers and forms that are usually abbreviated in writing are written out in full.
E.g.
Transcribed Written
zero, oh for 0
one for 1
ten for 10
one hundred for 100
nineteen oh three for 1903
et cetera for etc
saint for St
okay for o.k. (see above)
missus for Mrs
mister for Mr
but miss for Miss
ms for Ms

 

10. Transcript Examples

What follows are some examples taken from two actual transcripts. These examples include all relevant markup symbols and referencing.

 

<WSC#DGB004:0100:Z1><{><[>lee</[> wanted to get away from a labour intensive<,> er economy to a brains intensive economy and he's done it

<WSC#DGB004:0105:HS><[>how do <.>you</.></[></{>

<WSC#DGB004:0110:HS>mm

<WSC#DGB004:0115:HS>how do you turn those literacy figures around so fast

<WSC#DGB004:0120:HS><{><[>do you</[> have to be quite tough minded and harsh

<WSC#DGB004:0125:Z1><[><.>well</.></[></{>

<WSC#DGB004:0130:Z1>i don't know that it's a very harsh society really

<WSC#DGB004:0135:Z1>the schools are perhaps<,> <drawls>more</drawls> disciplined and authoritarian than ours but singapore children don't seem wildly unhappy <laughs>to the casual eye</laughs>

<WSC#DGB004:0140:HS>so it's something that they have there as well

<WSC#DGB004:0145:HS>what can we learn from their success or is their success in a way irrelevant to OUR lifestyle and OUR attitudes

<WSC#DGB004:0150:Z1>no i think we can learn a great deal

<WSC#DGB004:0155:Z1>keep in mind that when i first went to singapore thirty years ago they had leprous beggars in the streets and most of the houses didn't have running water in old chinatown and so on and er <O>voc</O> there was wall to wall slums <WSC#DGB004:0160:Z1>now <&>3:00</&> everybody's um housed properly they certainly have running water and so on er <O>voc</O> and<,> ONE of the reasons for that i think is er very CLEAR sighted government er and economic policy

<WSC#DGB004:0165:Z1>er WE debate the virtues or otherwise of intervention

<WSC#DGB004:0170:Z1>um the <?>p a p</?> government in singapore has certainly regarded one of their roles as being to direct the economy to encourage investment from outside to make er<,,> decisions about the TYPE of economy they want to encourage for instance er in the nineteen sixties they DOUBLED at a stroke the minimum wage because lee didn't want to persist with high um labour intensive <.>int</.> industries

<WSC#DGB004:0175:Z1>he wanted to FORCE employers to move across to <.>cap</.> capital intensive industries that sort of thing

<WSC#DGB004:0180:HS>so it's an enormous social experiment that's worked or social <laughs>laboratory if you want</laughs><latch> <&>4:00</&>

<WSC#DGB004:0185:Z1>well yes i mean there's some people would say <.>th</.> that the cost has been too high …

 

<WSC#MUC002:0115:JM>the ball's won by the all blacks

<WSC#MUC002:0120:JM>innes gets it kicks away with his left foot down <.>this</.> in his right hand touch line

<WSC#MUC002:0125:JM>martin has to scamper back not really scampering and the <drawls>ball</drawls> stops JUST over the goal line and he's forcing down in front of<,> john kirwan<,,>

<WSC#MUC002:0130:JM>and sean fitzpatrick

<WSC#MUC002:0135:JM>fitzpatrick signals to someone to get up and MARK where he should be on the twenty two for the dropout<,>

<WSC#MUC002:0140:JM>brewer goes in there standing on the line fitzpatrick now gets back<,,>

<WSC#MUC002:0145:JM>lynagh kicking to his left to this side of the field<,>

<WSC#MUC002:0150:JM>tapped down by gavin

<WSC#MUC002:0155:JM>on the hands of daley to farrjones

<WSC#MUC002:0160:JM>back <.>to</.><,,> who's that playing at first five cornish and cornish kicks but HIS touch kick hasn't found touch

<WSC#MUC002:0165:JM>kirwan knocks it on inside the half<,,>

<WSC#MUC002:0170:WS>yeah it's just a lack of concentration there isn't it

<WSC#MUC002:0175:WS>they've GOT to concentrate on the pass on just catching passing taking high balls<,,> <&>2:00</&>

<WSC#MUC002:0180:WS>very inaccurate this er australian back line aren't they with <{1><[1>their</[1> passing

<WSC#MUC002:0185:JM><[1>mm</[1></{1>

<WSC#MUC002:0190:WS>just ineffective at the <{2><[2>moment</[2> compared to the all blacks

<WSC#MUC002:0195:JM><[2>yeah</[2></{2>

<WSC#MUC002:0200:JM>i don't know what they've been doing at training cos we haven't been able to see it

<WSC#MUC002:0205:JM>but um i'm sure they've been SHARPER than this<,>

<WSC#MUC002:0210:WS>in comparison little and innes really in big games aren't they <{><[>mcbeth</[>

<WSC#MUC002:0215:JM><[><O>tut</O> oh yeah</[></{> yep<,> totally agree

<WSC#MUC002:0220:WS>i like terry wright's position here he's in nice and close <.>ther</.> <O>voc</O>

<WSC#MUC002:0225:WS>he's in a good position to arc out onto that wing and there're <O>voc</O> no options out wide for the aussies<,>

<WSC#MUC002:0230:WS>they've really got a good blanket defence this all black team today

<WSC#MUC002:0235:JM>well they've certainly they've worked on it wayne and er and it's paid dividends hasn't it hasn't it <{><[>but er</[> you know <.>you</.> you expressed er concern about it but they've sharpened up so much

<WSC#MUC002:0230:WS><[>for sure</[></{>