In order to analyze the speeches, a Content Analysis (CA) (Babbie, 2008;
Bauer, 2000; Rosengren, 1981; Weber, 1990) approach was used. CA unites statistical and qualitative analysis, and is therefore ideally suitable for the present study. It permits uniting theory based categories with the empirical findings in the texts themselves (Bauer, 2000).
There are two fundamentally different ways of analyzing natural speech:
one is to treat it as if it were merely a string of words, coding key words whenever and wherever they occur, subsequently ascribing them to the differing value categories. The other one consists in including in the analysis also the context, and taking into account the subtle underlying meanings conveyed. For this study the latter approach was used. 139
In natural speech word-count only would not give a true picture, as there are also many indirect ways of conveying meaning and referring to concepts, and therefore the focus could not be on merely nouns (e.g. Bardi, 2008;
Helkama & Seppälä, 2006) or on adjectives140, but also had to take into account the larger linguistic context. This also has to do with the nature of the materials. Firstly, the corpus consists of materials written in two very different languages, Swedish and Finnish. Till 1862 (the year when the language statute limiting the use of Finnish to religious and economic texts was revoked) all the texts are in Swedish and only after that in Finnish (see Sjöblom, 1995). Belonging to two different groups of languages, the way sentences (and hence meanings) are constructed is also different, making mere word count ineffective. Furthermore, due to the nature of the Finnish language and its agglutinative mode of word formation, there simply are too many variants of any given word for it to be meaningful to do a word count based analysis.
This was then also the rationale for using the Atlas.ti software rather than using other word-search based computer-aided coding software, such as e.g.
Alceste. Atlas.ti software was developed in Germany in the 1990s, and it has been widely used for computer aided qualitative data analysis. While its
139 See also Suedfeld & Weiszbeck, 2004.
140 For a more thorough discussion on linguistics and values, see Rohan (2000).
modus operandi is rooted in Grounded Theory, it is also very useful for CA, which unites qualitative and quantitative analysis of the data (Bauer, 2000;
Weber 1990). The software was used for the first two phases (see below). In the final phase coding was done manually.
6.1.1 CODES
Three categories of codes were used: firstly value codes; secondly, sc. anchor codes to anchor the texts historically and to aid the analysis; thirdly, text based codes.
6.1.1.1 Value codes
Value codes were ascribed using as codes the single value items and their dictionary equivalents provided by the Schwartz Value Theory (SVT)141, together with Work – related value items and Spirituality/Inner life value items, found in previous Finnish studies (e.g. Helkama, 1999; Helkama &
Seppälä 2006; Jaari, 2004; Myyry & Helkama, 2001; Verkasalo, 1996).
Following the definition of values (Schwartz, 1992)142, an expression was coded to be a value expression when it denoted desirability or importance.
The following examples clarify the coding principle, with the words guiding the coding underlined.
a) Sentence coded with Self-Direction as a value (single item:
independence, freedom)143:
Ne velvollisuudet, joita itsenäisyyden mukana olemme saaneet, kohdistuvat ensi sijassa itsenäisyytemme ja vapautemme säilyttämiseen ja
vahvistamiseen.
The duties, given to us together with our independence, aim foremost at the preservation and strengthening of our independence and freedom.
Kekkonen, NYS 1967
b) Sentence coded as not containing a Self-Direction value: the word
‘independence’ used to describe a timeline:144
Tarkastelkaamme teollisuustoiminnan kehitystä itsenäisyytemme aikana.
Let us look at the industrial development during our independence.
Kekkonen, NYS 1967
141 see chapter 2 on the theoretical background
142 Values are general, trans-situational abstract goals that guide both individuals and larger entities in the selection of actions and evaluation of events. These goals convey what is perceived as important or desirable.
143 Translation by the author
144 Translation by the author
The following two are examples145 of one SVT- based coded sentence (Benevolence, a SVT value) and one example of a Non-SVT value, namely Work–related values.
Benevolence
Inga medel hafwa derföre af Oss blifwit sparade, inga tillfällen att hjelpa och understödja förbisedde.
Therefore no means of help have been by Us spared, nooccasion to give help and support has been left unused.
Nicholas I, PDD 1832
Work-related values
Ennen kaikkea riippuu taloudellinenkin tulevaisuutemme kuitenkin itsestämme: sitkeästä ja tarmokkaasta työstämme. Vain kova työ ja lujat otteet voivat taloutemme pelastaa. Siihen perustuu elämämme ja tulevaisuutemme.
Above all our financial future too depends on ourselves: on our tenacious and vigorous work. Only hard work and firm actions can save our economy.
That [work] is the foundation of our lives and our future.
Paasikivi, NYS 1947
The essence of values is desirability, but in natural speech people also talk about what they consider to be threats to values. These were coded as well.
Thus, in the following example the sentence was coded as Universalism146.
[…]suursodan uhka on väistynyt Euroopasta, mutta rauhaa nakertavat alueelliset konfliktit , joista monien ratkaisemisessa kansainvälinen yhteisö on ollut lähes voimaton.
[…] the threat of large scale war in Europe has been averted, but peace is still been chipped away by local conflicts, many of which the international community has been unable to solve.
Ahtisaari, NYS 1995
Furthermore, based on previous studies (Aavik & Allik, 2006; Cacioppo &
Berntson, 1994, Cacioppo et al, 1997), negative value expressions were also included in the coding. The definition used for negative value expressions was simply ‘something to be avoided or something undesirable’. Thus an expression of a value, e.g. Hedonism was coded as negative, when the pursuit of the value was itself labeled as something negative.
Sammumaton on ihmisen onnen jano. Kuinka tosiaan nykyinenkin aika ponnisteleikse onnea saavuttaakseen! Ja kuinka monesti turhaan, syystä että onnea haetaan aistimaailmasta ja sen nautinnoista!
The thirst for happiness of man is unquenchable. How even the present times strive to obtain happiness! And how often it is in vain, because happiness is
145 idem
146 idem
sought from the gratification of sensuous pleasures!
Senate, PDD 1897
The following categories emerged from the texts: negatively valued Hedonism, negatively valued Tradition, negatively valued Power, negatively valued Achievement, negatively valued Stimulation Negative, and negatively valued Self-Direction. In the final analysis they were dropped due to their low frequency, and because they did not correlate with other values in the model, therefore suggesting they might not be part of the same value structure as their positive counterparts.
6.1.1.2 Anchor codes
The following codes were used in order to anchor the texts to their historical and political context and in order to permit further quantitative analysis:
1. Speech type:
POS for Parliament Opening Speech, PDD for Prayer Day Declarations, NYS for New Year’s Speech, OTH for other public speech.
2. Time-related codes: century, decade, and year.
3.Ruler: Czar, President, Senate, other (such as PM). These were broken down further into individual name-based codes These anchoring codes were then also annotated with relevant information to allow contextualizing of particular expressions, thus enhancing the depth of the qualitative analysis of the texts themselves.
6.1.1.3 Text-based codes
Text based codes were codes directly arising from the text such as e.g. the addressees of the speeches.
The addressees of the speeches were coded looking at the explicit expressions. Where the addressees were referred to more implicitly, this was coded under “all people”. The latter code was created to allow for the usage of this as a device creating boundaries for in/out group formation, and for the inclusiveness of the moral universe (see Schwartz, 2007).
The following two quotes illustrate the difference; the first was coded as
“addressees (subcategory Lutheran Finnish subjects)”, the second example was coded under “all people (subcategory Finns)” and “all people (subcategory citizens)”:
1.
[…]Tillbjude Eder, OSS älskelige WÅRE trogne undersåtare af Lutherska församlingen i Stor Furstendömet Finland, WÅR synnerliga ynnest, Nådiga benägenhet ock gungliga wilja, med Gud Allsmägtig!
[… ]Offer You, OUR beloved unto us faithful subjects which are of Lutheran congregation in the Grand Duchy of Finland, OUR special favour, gracious predisposition and Our good will with God Almighty
Alexander I, Prayer Day Declaration, 1812
2.
Meidän suomalaisten on hoidettava maatamme niin, että se säilyy itsenäisenä ja vapaana, ja niin, että kaikki kansalaiset voivat katsoa luottavaisesti tulevaisuuteen.
We Finns have to take care of our country in such a way that it remains independent and free and in such a way that all citizens can look to the future with confidence.
Ahtisaari, New Year’s Speech, 1995
6.1.1.4 Coding unit: sentence
The texts were coded at sentence level, except for some of the 19th century texts, in which there were sentences with many subordinate clauses, in which case they were coded at clause level. However, in natural speech the thought is often carried over from one sentence, and therefore the sentences were never looked at in isolation. The sentences were coded with as many values as they contained, more than one if necessary. The following excerpt illustrates the point.
Inhimillisen kulttuurin ja henkisen sivistyksen työsaralla kansamme velvollisuudet ja mahdollisuudet näyttävät yhä suurenevan. Tämä osallemme lankeava jalo ja korkea tehtävä kannustakoon meitä erikoisiin ponnistuksiin.
It seems that in the area of human culture and inner civilization the duties and possibilities as a people seem to constantly growing. May this high and noble task, which we have been given, challenge us to extraordinary efforts.
Svinhufvud, New Year’s Speech, 1937
This passage was coded as BE (duty), ST (possibilities), ED (human culture and intellectual/inner civilization147), and WO (exhortation to endeavor).
One of the difficulties in ascribing codes was that in natural language, the combinations ‘main word+attribute’ often yield value combinations that are not motivationally close. For example, the following sentence can be coded as Achievement, or Security, or Self-Direction.
Olemme hakemassa paikkaamme Euroopassa. Kysymys on […]
mahdollisuudesta turvata omat etumme […]
147 Education values emerged from the initial Spirituality value category, see section 6.2 on The next phase below
We are looking for our place in Europe. It is a question of […] possibility to safeguard and secure our own interests, what is beneficial to us […]148 Ahtisaari, Parliament Opening Speech, 1995
The following example shows a similar problem where the first part is part of Education values, and the second Security and Power values.
Maamme ja kansamme kehitys vaatii ehdottomasti säännöllisten taloudellisten olojen aikaansaamista
The progress and development of our nation and people absolutely requires that we bring about regular economic conditions149
Mannerheim, Parliament Opening Speech, 1919
These were also the cases in which the inter-rater agreement was often reached through discussion, if the raters had coded these instances differently. In the examples above the larger linguistic context was also taken into account, and then the sentence was coded with all values which were present.