This study was conducted using the Eighteenth Century Collections Online (ECCO), which contains over 180,000 titles (200,000 volumes) and more than 32 million pages, making it the premier and irreplaceable resource for eighteenth-century research. ECCO was created on the base of the English Short Title Catalogue (ESTC) in 1994 after expanding it to other periods.4 This metadata catalog currently includes bibliographic information for over 480,000 documents that were published between 1473 and 1800. Most of the documents included in the ESTC (and also in ECCO) were written in English and published in the British Isles and North America, collected from hundreds of libraries worldwide in a union catalog coordinated by the British Library. After the ESTC had been launched, “the Eighteenth Century” microfilming project followed in 1981. ECCO was born when microfilms of that project—originating mainly from the British Library, Oxford, and Cambridge—were scanned in 2000–2002, and optical character recognition (OCR) was applied to associate the pages with automated transcriptions of the texts. These texts were published in a web interface in 2003 that also enabled text search of the transcriptions (Tolonen M., Mäkelä E., Lahti L., 2022).
In this study we chose to focus on the political pamphlets published in England in the first decade of the 18th century. Further in the bias it is mentioned that political pamphlets of the beginning of the 18th century are better represented in ECCO than other decades, and the reason might be that not all the pamphlets made it to the collection. Since the pamphlets were one of the main sources of the information of the political and social changes and governmental decisions of the mentioned period, we assumed that they have a strong potential for reflecting the political battles. The goal of pamphlets was to deliver the information to more than one person, unlike personal letters, instead they were read to a large number of people. This tells us about the high importance and the influence pamphlets had which can be equal to the media sources nowadays.
Taking into consideration the importance and influential potential of pamphlets as well as their sufficient representative amount in ECCO we chose to carry out the research on the political pamphlets of the first decade of the 18th century.
Definition of Pamphlets
Various definitions have been made to define political pamphlets throughout time. Verhoest (2018), defined pamphlets as “short tracts of religious, political or topical interest, written in the vernacular, generally printed in a quarto format, and sold at a low price in order to reach a wide audience” and were the only most significant print medium in England and many other countries in the late 17th century (p.48). The differences in format of the pamphlets resulted in different classifications such as “quarto, octavo, and folio” (p.5) and due to these changes Raymond (2006) defined typical pamphlets’ length as one page to twelve pages while the ones in octavo format as between eight pages to ninety-six pages.
Importance of Political Pamphlets
Pamphlets were published to inform people about political issues, decisions, and recent events. In addition, they were sold by means of profit and persuasion and communication. The significance of pamphlets as a regular feature of bookstores and the economy of book trade after the late 16th century increased (Raymond, 2006). Moreover, pamphlets were one of the most printed forms in the 1700s. Hence, they had a big audience and “by 1700 everyone knew what a pamphlet was and what it did” (Raymond, 2006, p. 7). They drew enormous attention from all parts of the public. Thus, they were read not by a single person, but to a large number of people. The pamphlets reflected the main political battles of those times. They had power to affect political events, shape public opinion, and create a discussion among individuals which turned out a popular debate (Raymond, 2006). Using pamphlets as a political tool to reach a larger and less restricted audience increased after the Commons agreed to publish the Grand Remonstrance (Raymond, 2006) which resulted in multiplied pamphlets that responded to one another similar to the dialogue form.
Bias and representation
When determining the way pamphlets could represent the political environment, it is necessary to consider the possible bias and the representation coverage of the pamphlets in ECCO. Even though ECCO is a great example of digital collection archive, which provides access to eighteenth-century texts in the English language, we need to remain critical about the overall representativeness of the data. Smaller books may be more easily lost over time, general collections are often complemented with specialty collections, and certain types of information such as that about the publisher and the author may be more likely to be missing, or more challenging to interpret, for certain time periods (Tolonen M., Mäkelä E., Lahti L., 2022).
Since it is impossible to give a precise definition of an eighteenth-century pamphlet or book, because neither of them is a generic category, we defined them as a maximum 32 pages for a pamphlet-sized document, and a minimum 128 pages for a book, regardless of the leaf size (Tolonen M., Mäkelä E., Lahti L., 2022). When counting the total amount and the representation of the pamphlets during the decades, we can see that the amount of pamphlets was declining through the eighteenth century (Tolonen M., Mäkelä E., Lahti L., 2022).
(Tolonen M., Mäkelä E., Lahti L., 2022)
Taking into consideration that the total amount of printed books was increasing, not all the pamphlets made it into ECCO. The first decades are much better represented, which was one of the factors of choosing the time period in the current research. At the same time we need to keep in mind the fact that even if the collection of the early 17th century is considered good, it might not include all the pamphlets issued in the mentioned period. Nevertheless, the amount of the pamphlets and text snippets found allow us to see the patterns in political life.
- What are the most popular text snippets from 1700-1710 in political pamphlets?
- How do these results represent the political environment of the decade?
- How do the findings relate to the political-religious discourse of the 18th century?
- Do certain text snippets predict (ecco) genre?
The pre-existing genres in the Ecco database were
- Fine Arts
- General Reference
- History and Geography
- Literature and Language
- Medicine, Science and Technology
- Religion and Philosophy
- Social Sciences.
Out of these genres, we figured Law and Social Science would include political texts. We defined political pamphlets in the following way:
- text snippet length < 32 pages
- genre = Social Science OR Law
- publication year < 1711
Our final data included 3000 pamphlets and 1.9 million text snippets.
We used two different methods to count the top 500 text snippets and top 200 text snippets, where length of the text is over 200 words.
Even though a text snippet borrowed implicitly exists in the text reuse data, there is no such thing as a text snippet id in the data. Instead, a single reuse item consists of the text ids of the two matched texts, and the start and ending points in those texts (t1_id, t1_start, t1_end, t2_id, t2_start, t2_end) and some more. This meant that for our project, we had to decide which reuse items represent the same text being reused. The data doesn’t contain the text itself so the only way to find out which reuse items represent the reuse of the same snippet is to compare the reuse items. If two items are similar enough, we say they concern the same snippet. Because comparisons are between two reuse items at a time, we’d like to have some properties hold no matter how we end up defining similarity. To define these properties, let’s define xRy to mean “x and y satisfy the relation R”. Now we can write our desired properties out as such:
- xRx for all x (reflexive)
- xRy => yRx (commutative)
- if xRy and yRz, then xRz (transitive)
A relation with all these properties is called an equivalence relation. If these properties hold, then one only needs to choose a representative of a text snippet and compare it to other reuse items to find every reuse that concerns the same snippet. However, this is not necessarily possible. Consider for example, the following case. Let’s note strings of arbitrary length greater than zero with lowercase letters. Text A is “abcd“, text B is “xbcy“, and text C is “fbcd“. Now the relevant reuse items would indicate that the matches are as follows
A <-bc-> B
B <-bc-> C,
A <-bc-> C. Instead we have A <-bcd-> C, and as such the desired transitivity cannot be a property of our equivalence relation.
The following algorithm was used to create a partition of our subset of text reuses.
1. pick the first reuse item in the list
2. find every other reuse item where
a) t1 matches
b) t1_start matches (within +-10% of align length )
c) t1_end matches (within +-10% of align length)
3. mark all of those with the id of the reuse item
4. remove all matches and the reuse item from the list
5. repeat until the list of unprocessed items is exhausted
This is only good for an approximation of how popular a particular snippet was, because it only concerns itself with reuses from the same text. This would be enough if transitivity was possible, as every snippet that is the same, even if the text it is in is different, could be expected to have the same number of reuses (with some small margin of error) for every text in the dataset that uses it. Without transitivity, however, this result is only good for an approximation. Luckily the most widely-used snippets repeated in the top 500 so often that any margin of error this method produces most likely is not significant.
Another problem with this algorithm is that due to the selection of margins for matching the start and end of a snippet, it is not necessarily even commutative. Once again, this is inconsequential due to the reasons mentioned above.
For future research, we ran a similar analysis for situations where t2 entries match, and where t2 of one matches the t1 of another, marking every reuse item with the tr_ids of those reuses that it matched with. Unfortunately, we did not have the time to analyze the results. The next step would be to create a partition of snippets by starting with a reuse item and then using breadth first search to find all other items related to it, and then taking the next one until there are no unprocessed items left. The data is ready and a correct algorithm has been tested but also unfortunately deemed slower than doing things manually.
In this approach, we would pick M times the most common t1_id, which is the text id of the text that has been reused in the other text (t2) in the pair (or in one row). For each t1_id, we pick N times the most common t1_start, which is the starting point of the snippet in the pamphlet, and t1_end, which is the ending point. T1_start and t1_end are checked separately, because there could be text reuses that always end at the same point, but where the reuse starts varies, and vice versa.
For the N most common t1_starts and t1_ends, we get the corresponding most common t1_end for the t1_start and t1_start for the t1_end. The number of times these pairs appear in the whole pamphlet data will determine the top order. When counting the occurrences of a start and end pair, two starts and two ends are considered to be the same, if the difference between the starts and the ends are 10 % of the text length of the original pair at max. If both start and end match, an occurrence is counted. After counting all occurrences of one pair, the results are saved into a dictionary. The dictionary is sorted by the count of pairs after going through all pairs of all t1_ids.
We tried different Ms and Ns, but the final results we got when M = 650 and N = 10. These numbers were completely arbitrary, not much comparison with other numbers were done. M is 650 so we could get well over 500 results to sort, and N is 10, because we thought one pamphlet probably won’t have more than ten snippets of text that are reused more than the next most common t1s’ most common text snippets. The same method was used to calculate top the 200 longer reuses, and only the dataframe of all pamphlets was filtered so that it included the rows where text length was over 200.
Comparing the results
Especially both of our top 500 reuses were contextually very similar, if not practically the same. The top 200 longer results were also similar, but varied more. Interestingly, despite the different approaches, we found similar patterns in our results. In the scatter plots below are first both top 500 results, and under both top 200 results. On the x-axis is the top order and on the y-axis is the t1 ids.
Unfortunately we didn’t have time to study the similarities further, but it would be interesting to identify those pamphlets that seem to appear several times and study the reason behind it.
As the genres (ecco_module column in ecco_core table) of our pamphlets are Law and Social Science, we were interested in the genre distribution of the pamphlets’ t2 texts (text where a snippet (t1) is reused). This was a pretty simple study after we figured out the SQL-query for getting the genre of the t2 texts. As seen in the charts below, we calculated the distributions for Law and Social Science separately.
We were happy to see that for both Law and Social Science, the most common t2 genre was the t1 genre. For law, this percentage was 64,2 % and for Social Science 58 %. In addition, in both cases, the second most common genre was the other genre of these two. This gave us confidence that the genre of a text snippet is actually Law or Social Science, if the t1 – t2 pair has the same genre.
Here we found out that t1 genre can predict t2 genre for Law and Social Science. One thing to study further, which we didn’t have time for in this course, is to whether single text snippets can predict genre or not. Probably this isn’t equally distributed between the snippets. If we could find that a certain text snippet or sentence in a context is about some topic, that would be quite a big deal. This could be hard even for a human, so it would be prominent to have a computer do that correctly.
Case Study 1: Dissenting Anglicanism or causing chaos? Representations of the Clarendon code throughout the 18th century
Introduction and previous research
(1) “I, A. B. Do swear, that it is not lawful upon any pretence whatsoever, to take arms against the king. And that I abhor that traitorous position of taking arms by his authority against his person, or against those that are commissioned by him, in pursuance of such commissions.” (Humfrey, 1702: 4)
In this case study, the text reuses of the snippet from the collection of legislation known as the clarendon code, shown above in example 1, will be examined to form a tentative understanding of how religious-political discourse manifested throughout the 18th century.
The text reuse snippet was chosen from the top 200 long text reuses (that is, texts with more than 200 characters) of political pamphlets. This threshold was shown by both Lari’s and Inkeri’s method to greatly favour texts detected as law (hereafter referred to as the legal genre) in t2. Text reuses that then are distinctly political instead of legalistic within this subset, should provide an analytically interesting, anomalous starting point for this study.
Using the text reuses of the clarendon code (hereafter referred to as CC) as a resource for analysis, I hope to illustrate how dissenters and nonconformity are represented throughout 18th century political literature and what that can tell us about the broader political environment of the time. Nonconformity and dissenters, especially in the 17th and 18th centuries, were affiliated with political radicalism and by some critical voices condemned for their “groundless discontent” and aims for “universal anarchy” (Bradley, 1990: 158). Aims to limit the practices of dissenters, who were increasingly being seen as a political threat to the monarchy, resulted in the passing of the collection of legislation known as the Clarendon code, starting with the corporation act in 1661 and followed by the act of uniformity in 1662 (of which the text reuse snippet used for this study is from), the conventicle act of 1664 and, finally, the five mile act of 1665 (Dudley, 1912: 69).
Nonconformism in 18th century England became increasingly aligned with the “party of progress” (Bradley, 1995: 5), the Whigs: a natural ally to nonconformists who shared with them, the ideals espoused by liberal thinkers such as John Locke (Bradley, 1990: 158). Furthermore, the first half of 18th century English politics is marked by a so-called “Whig supremacy” whereby the Whigs achieved complete dominance over the government, ousting representatives of the Tory faction from all significant positions of power (Miller, 1998: 73). Therefore, it should prove interesting to discover how the findings of this case study reflect such a time of Whig supremacy. Will sympathy for the nonconformists be reflected in the attitudes towards CC, or will support for the act mirror the dominant Anglican tradition of the time?
Distribution of text varieties
To examine how religious-political discourse takes shape in 18th century literature as it pertains to CC, we should first identify what types of genres are represented in the data.
From the data, I first identified 217 yearly unique texts (hereafter referred to as YUT). YUT refers to those reuses of the snippet from unique works as identified by Rahti. So, in this category no multiple publications of the same work (published within the same year) or multiple text reuses from the same work were included. Argumentative texts (hereafter referred to as AT) refers to those unique texts that were identified through the process of multiple close readings as advocating or arguing for or against CC. In this subset, texts originally written before the 1700s are not included to better gauge the perspectives proposed specifically in the 18th century. The close reading of the 217 YUTs identified 62 ATs, which will be further examined in sections 2.2 and 3 below.
In this section, I will first identify the genres present in YUTs and then move on the ATs and their genre distribution to hopefully gain some primitive understanding of the types of 18th century literature that were interested in CC.
2.1 Distribution of text varieties within yearly unique texts
As can be seen in figure 1 below, the 217 YUTs fall into four broad text varieties: historical, political, religious and legal genres.
Texts included in the historical genre are mainly histories of England or the United Kingdom or histories of an important figure. Additionally, accounts of parliamentary proceedings are also represented in this text variety. 66 texts in total were identified to belong to the historical genre. Texts in the political genre are outright argumentative in tone, often author-based in style and typically advocate for differing political issues. All in all, 48 texts were identified to belong to the political genre. Texts in the religious genre consist mostly of renditions of the Common Book of Prayer, a book narrating the proper religious practices as it pertains to the church of England. Additionally, the religious genre further includes practical sermons, or texts that itemise religious justifications and perspectives on matters in a more argumentative tone. In total, texts belonging to the religious genre amount to 49 texts. Finally, the legal genre consists of descriptive texts reciting the legislation of the time and represent roughly a fourth of all the texts identified in the process of close reading, amounting to 54 texts in total.
Figure 1. Distribution of text varieties within YUTs
Therefore, although the distribution of the four genres is relatively even, we see that the texts that fall into the category of political literature, represent a minority within the text reuse data. Considering this, we could argue that overt political stance-taking in literature pertaining to CC remains rare in comparison to the covert stance-taking veiled by a religious doctrine for example (for more on this see section 3.1) or the descriptive approaches as seen in much of the historical and legal genres.
2.2 Distribution of text varieties within argumentative texts
To then look into how the argumentation surrounding CC is structured in 18th century literature, we must first identify from the 217 YUTs those texts that assume an argumentative stance.
As represented in figure 2 below, argumentative texts, though mostly consisting of texts from the political genre (40 texts in total), also include works in both religious (6 texts) and historical genres (16 texts). This is due to the presence of practical sermons in the religious genre, and critical histories in the historical genre. Both text types argue for or against CC, either on the basis of religious doctrine, or critique against the predominant political order. However, no texts from the legal genre were included in this subset, as their content was found to be solely descriptive in nature.
Readers with a keen eye might notice that although 48 texts were identified as part of the political genre in YUTs, not all political literature is included in this patch of argumentative text data. This is because although all texts within the political text variety were argumentative in tone, not all were originally written in the 18th century. For example, the database includes multiple collections of John Locke’s writing republished throughout the 1700s, even though Locke passed away in 1704. This distinction is key for the following rhetorical analysis on CC and its perception particularly throughout the 18th century and therefore those eight texts that were not originally written during the century had to be discarded from further analysis.
Figure 2. Distribution of text varieties within ATs
In this section, I will further examine the ATs within their textual and political realm, to gauge how nonconformity was perceived as a religious and cultural force throughout the 18th century.
In figure 3 below, we can see that from the 62 ATs, a majority of the texts assume an opposing position to CC. All in all, while 26 texts support the legislation, 36 texts were in opposition to it. Considering the limited amount of data, it is difficult to draw any broad implications from this. For example, some modern day data collection bias towards on certain kinds of works on ECCO’s part may no doubt skew the results one way or the other.
Figure 4. Distribution of support and opposition of CC in ATs
Nevertheless, these results are in contrast with the contemporary understanding of liberalism as represented by the nonconformists and the Whig party as the “dominant ideology of opposition” to the majority ideology of Anglicanism (Bradley, 1990: 158). This could be explained by the fact current day data collection and digitization processes favour the more liberal position proposed in the articles opposed to CC. Or, alternatively, it may very well be that CC itself was deemed so uniquely reprehensible a bill, that support for it is diminished, even though in other areas of polity an Anglican position prevails.
3.1 Anglicanism, divine authority and passive obedience
Those in support of CC, typically argue for the act on the basis of two predominant ideologies and doctrines of the time: conservative Anglicanism and passive obedience.
What emerges from the texts in support of the bill, is the Anglican ideal for “obedience … to the government and the need for a stable social order” (Bradley, 1990: 158). Dissenters were described as “schismatics” or separatists from the church of England (Astell, 1704: 91) and their political agenda was seen to be identified first and foremost by factionalism and instability.
In support of CC, a conservative counter-argument to the liberalist one proposed by the nonconformists is then formed, whereby change of any kind (be it within the church of England, the government or the monarchy) is condemned an approached with suspicion. In the following example, Astell describes this anxiety of the chaos caused by the change proposed by the dissenters:
(2) “Innovations are seldom for the better, wise and good men had rather bear some known inconveniences, than run the hazard of greater by a change, which usually brings such as no humane prudence could foresee or provide against.” (Astell, 1704: 90)
This very anxiety is then argued for from several angles. For example, a subgenre of practical sermons emerges from the religious genre. Sermons, alongside pamphlets, were the most widely used methods of political messaging and influencing throughout the 18th century (Bradley, 1990: 124). And while most sermons were simply orated by a preacher, some sermons were additionally published in text form. The 6 ATs from the religious genre all represent the subgenre of practical sermons and are also all in favour of CC.
Within the practical sermons, a doctrine of passive obedience as an argument for CC seems a prevalent feature. Passive obedience is a religious and political doctrine of absolute submission to law and authority, and a prevalent political ideology of the time whereby disobedience to an authority of any kind is disagreeable. To illustrate the severity of passive obedience, Berkley for example writes that, even though tyrants “are the worst and vilest of men, and … have not the least right to our obedience. But the laws of god and nature must be obeyed, and our obedience to them is never more acceptable and sincere, than when it exposeth us to temporal calamities” (Berkeley, 1712: 31). In the following example, the doctrines of passive obedience and the sin of resistance are used in support of CC:
(3) “Q. Very Well: Let me hear your proofs that resistance is against the laws and constitution of the kingdom.
A. Yes: […] [CC]
Q. Very well: But may not subjects resist, when opprest in their establish’d religion?
Q. How do you prove it?
A. By scripture, the doctrine of the church, and the constitution of the kingdom.” (“The loyal catechism”, 1713: 12)
In the same text, context as to why such resistance is to be prohibited is also given:
(4) “Indeed, a rebel is worse than the worst prince, and a rebellion is worse than the worst government of the worst prince that hitherto hath been.” (“The loyal catechism”, 1713: 16)
Therefore, we can observe how the conservative argument for CC seems to be tied with the Christian doctrine of passive obedience. That is, both arguments seem to share a quality of anxiety about change and the chaos caused by rebellion and resistance. Therefore, CC, that limited the dissenters’ opportunity to practice Christianity within their parameters, is then considered a just collection of legislation.
3.2 Nonconformism and liberal thought
In contrast, if we look at how CC is described by those in opposition to it, we see an emphasis put on the severity of the laws as a particular point of contention:
(5) “It would be easy to fill a volume with the worthy exploits of this sort perform’d by this parliament: how stoutly they suppress’d all conventicles: […] abolishing the rights and libertys of the people, and erecting in their stead passive obedience and non-resistance” (Morris, 1741: 68)
Particularly, the aforementioned values of an individual’s rights and liberties seem to be a recurring feature in much of the ATs positioned against CC. Although we must acknowledge that nonconformity and liberalism (especially as we know it today) are by no means synonymous and, in fact, share a multitude of contradictions with one another, within the context of CC, an undeniably liberal argument is being advanced. A fear of disorder, much like the one present in those texts supportive of CC, emerges here too. However, unlike the Anglican fear of anarchy and chaos, what we observe in the dissenting texts is a fear of loss of one’s liberties:
(6) “From hence the crown derived so great an advantage that for the greater part of the last century our liberties were brought into continual hazard: it was the revolution alone, and the principles established then, that could have possibly preserved us.” (Perceval, 1743: 149)
Similar fears awoken by the legislative power of CC emerges in other texts weighing the implications of such legislation within the 18th century political context:
(6) “This shock to the national constitution is not to be over-looked. It is yours, to lay it before the parlement. For, if the parlement can vest a legislative power in any body, what avails our elections?” (Lucas, 1749: 18)
So, while ATs from the Anglican point of view seem to emphasise submission to authority to guarantee political stability, dissenting texts seem to condemn authority (in this case the monarchy and the government) as the source of said instability.
In conclusion, what we can determine from the results of this case study is that, while texts from the political genre dominate the field of argumentative texts concerned with the Clarendon code, both historical and religious genres are represented in the political discourse of 18th century literature by taking an active and argumentative stance on CC.
While a majority of ATs examined in this case study were in opposition to CC, we also find that support for the legislation remains strong throughout the century as well. Additionally, the argumentation present in the data is heavily influenced by religious doctrines, such as that of passive obedience, making the separation of the strictly political, and strictly religious within the 18th century context challenging. The conservative fear of disorder and political instability overlaps with the anxieties of rebellion the Christian doctrine of passive obedience aims to prune
Future research within this field could further examine how and in what light popular dissenters, such as Edmund Calamy, and their writing is represented and reused through time. Additionally, comparative analysis between representatives of dissenter and Anglican ideologies could further shed light on how these movements were viewed throughout the 18th century and how political discourse around them manifested.
Case Study 2: “Lords Spiritual and Temporal”
In our case study 2, we have examined the most reused 200 snippets to answer two of our research questions which are “What are the most popular text snippets from 1700-1710 in political pamphlets?” and “How do these results represent the political environment of the decade?”. We found that the most common reuse was “Lords Spiritual and Temporal” whose abbreviation is “LS&T” which we will use from now on to refer to such reuse.
Most of the pamphlets which reused the text of “LS&T” that we examined had relevance in terms of form and structure of the publication. The first page of the pamphlets was a cover page that included “LS&T” as a heading and/or informative text that aimed to represent the group of people who intended to share their ideas with the audience and/or the Queen. In the most common form of the cover page, the order of the heading, sub-headings, and the publisher’s details were divided into three. The heading was usually typed in the format as follows;
Of the Right Honourable the
Lords Spiritual and Temporal
In Parliament Assembled,
The words were written by paying attention to the upper and lower cases and the font sizes of specific words were different from each other. This form was followed in the second and third parts of the cover page. The second page of the pamphlets that followed this design and publication format included a text snippet for the Queen’s answer to the letters that were addressed to her and/or a particular informative text about the topic of the pamphlet in a definite structure which reused “LS&T” text.
The reuse of “LS&T” text was frequently used in political pamphlets which were published to inform the audience about the political issues between the Lords Spiritual and Temporal and the House of Commons, both parties’ consultation to the Queen or their recommendations to the Queen, the Queen’s answers to them, and so on by sharing letter exchanges. The content of the examined political pamphlets which contained the text reuse of “LS&T” derived in terms of political events of the time. However, the predominant subject of the political pamphlets was the conflict between the Lords Spiritual and Temporal and the House of Commons which serves as a significant example of the reason for the manifold multiplied political pamphlets that were published in the era and the increase in pamphlets being a political battlefield (Verhoest, 2018).
As may be expected the reuse of “LS&T” in a text increased at some points according to the length of the pamphlets whose length varied in terms of the topic, addresses, and form of the pamphlet. Lords Spiritual and Temporal preferred to write relatively shorter letters while thanking, supporting, and/or complimenting the Queen and throne so that they used their title fewer times when it is compared to the times that they addressed themselves as Lords Spiritual and Temporal while sharing their recommendation or complaints about the House of Commons in considerably lengthy letters with the Queen. Furthermore, the Lords Spiritual and Temporal attached precedents, whose number exceeded one at some parts which made the political pamphlets longer and provided a basis for more reuse of “LS&T”, to their letters or published them solely sometimes to present the relation of some previous incidents with the events that they encountered. Therefore, the “LS&T” text snippet was reused various times and became the most commonly reused text snippet of political pamphlets from 1700 to 1710.
To conclude, our case study 2 has met the needs of the research questions that we aimed to find answers to. By doing this case study, we examined the most popular reused text snippets from 1700 to 1710 in political pamphlets and encountered the political cases that represented the political environment of the decade. The most popular reused text snippet during the first decade of the 18th century vividly demonstrated the political activity of the time, the relationship between the Queen and the Lords Spiritual Temporal and/or the House of Commons.
Our group studied the most popular text reuses in political pamphlets from 1700 to 1710. We started from filtering political pamphlets from the Ecco data, based on our own definition of a political pamphlet. Two sets of top 500 text reuses and top 200 longer text reuses were calculated with two separate methods, and with close reading, we confirmed that the text snippets were from political pamphlets. We did two case studies. One of the most common text snippet and one of a text snippet we found interesting. For both of the texts, we were able to explain how the text snippets relate to the political environment of the time. Many aspects from our project work could be studied further, but we ran out of time. However, we are glad that we found an answer to all our research questions:
1. What are the most popular text snippets from 1700-1710 in political pamphlets?
“Lords Spiritual and Temporal”
2. How do these results represent the political environment of the decade?
The Queen consulted The Lords Spiritual and Temporal and The House of Commons and both of these parties exchanged ideas on political issues. Apart from those, the conflict between The Lords Spiritual and Temporal and The House of Commons about the political issues represent that they had different views on political issues of the time.
3. How do the findings relate to the political-religious discourse of the 18th century?
Political discourse emerges and propagates in not just political, but also historical and religious genres, with distinct styles of argumentation. The concepts of passive obedience and divine authority are used to argue for the Clarendon code and seem to share a quality of anxiety about change with other texts in support of the legislation. Within the limited amount of data, dissenting texts and opposing positions to the Clarendon code represent the majority position throughout 18th century argumentative literature.
4. Do certain text snippets predict (ecco) genre?
For genres Law and Social Science, there’s a ~60 % chance that t2 genre is the same as the t1 genre.
Reflection on learning
Inkeri: I calculated the top 500 most used text reuses and the top 200 most used long text reuses my own way (method 2). I also studied genre prediction and the similarity of method 1 and method 2 results. Overall, I think the biggest lesson for me was that sometimes a good guess goes a long way. Because of the structure of the data, we had to come up with ways we thought would get us the most common text reuses. During this project, I’ve gained more experience in working with imperfect data and the pandas library. It was nice to see how the data I had processed led to meaningful results with the expertise from another field of study, which is something I usually don’t get to see in courses with only other data scientists or computer scientists.
Miro: I worked on case study 1 to examine the political-religious discourse of 18th century literature and find answers to our RQ3. The greatest takeaway from this course was learning how to operate and examine such large amounts of, what one could call, “inconvenient”, messy data. Although I have my fair share of experience with corpus linguistics, typically those bodies of text have been fully digital (as in texts originally written digitally), making examinations of them simpler and the results cleaner in this regard.
In this course, I also gained a more in-depth understanding of the many ways in which digitalized data can be gathered, organized and examined for analysis and thereafter made meaningful within a broader context provided by ECCO’s huge databank and the text reuse methodology’s capability of finding relevant connections within this dataset. The concepts of text reuse and genre detection are new and exciting methodologies for me that I am eager to look into further and perhaps include in my own, future research. The possibilities of intermixing these two methodologies within future research seems like a promising avenue of helping researches understand the textual world of the 18th century in a more quantitatively satisfactory manner.
This course also made me appreciate the difficulty of achieving the meaningful data necessary for this type of research, thanks to the interdisciplinary nature of this course. Getting to see how students from different fields operate the data to gather meaningful results by using the methods mentioned above and hearing about the complications faced along the way was incredibly illuminating and something I will definitely take with me in my future studies.
Sevil: I worked on case study 2 and examined the political pamphlets in which the “Lords Spiritual and Temporal” text snippet was reused. I came across some popular political events of the time which helped me understand the political environment of the decade. Being able to read and evaluate one of the most popular communication tools of that time made me more enthusiastic to work on the project. Moreover, I enjoyed the process of close reading which was informative. In addition, I gained experience in the structure of the written discourse of pamphlets. I reckon that the project enriched my knowledge of texts and history. Apart from these, working with diligent teammates from different backgrounds assisted my learning and broadened my horizon. In conclusion, this project contributed a lot to me during my learning journey.
Lari: I was involved in deciding the research questions, and also worked on generating top 500 and top 200 most reused snippet lists. I also intended to generate actual snippet ids based on that work, but time ran out before I could finish that. When this course started running I’d done one little thing with pandas, but had no real knowledge of it. At this moment, I feel like I still have poor understanding of the system, even though I can now work with binary masks at the very least. The best learning experience I had was with formulating a real-life problem in mathematical terms, and then using that model to help think about the original problem.
Ekaterina: together with the team I was involved in the discussion and determining the research questions, definition of the pamphlets and in other parts of the research which refers to the humanities sphere. After the part of the team responsible for the data provided us the result, we analyzed the snippets and did a close reading of the most popular snippents. Based on them I also worked on the case study 2 about the Lords Spiritual and Temporal. The research gave a deep understanding of how projects in Digital Humanities held and how the cooperation with the specialists from Data science can be organized. Besides, the project gave me a great learning experience of teamwork with people of different backgrounds and mindsets, which can be used in future studies and career paths.
Raymond, J. (2006). Pamphlets and Pamphleteering in Early Modern Britain (Cambridge Studies in Early Modern British History).
Verhoest, P. (2018). Seventeenth-Century Pamphlets as Constituents of a Public Communications Space: A Historical Critique of Public Sphere Theory. Theory, Culture & Society. Volume 36, Issue 1, January 2019, pp. 47-62 https://doi.org/10.1177/0263276418779185
Vesanto, A., A. Nivala, H. Rantala, T. Salakoski, H. Salmi, F. Ginter. (2017).“Applying BLAST toText Reuse Detection in Finnish News-papers and Journals, 1771–1910.”In: (Proceedings) 21st Nordic Conference of Computational Linguistics. Gothenburg, Sweden, 23–24 May 2017 (Linköping 2017), http://www.ep.liu.se/ecp/133/010/ecp17133010.pdf.
Astell, M. (1704). Moderation truly stated. London: Printed by J.L.
Berkeley, G. (1712). Passive obedience, or, The Christian doctrine of not resisting the supreme power, proved and vindicated upon the principles of the law of nature: In a discourse deliver’d at the College-chapel (2nd ed.). London: Printed by H. Clements.
Bradley, J. (1990). Religion, Revolution and English Radicalism: Non-conformity in Eighteenth-Century Politics and Society. Cambridge: Cambridge University Press. doi:10.1017/CBO9780511560378
Bradley, J. (1975). Whigs and Nonconformists:” Slumbering Radicalism” in English Politics, 1739-89. Eighteenth-Century Studies, 9(1), 1-27.
Dudley, A. C. (1912). Nonconformity Under the” Clarendon Code”. The American Historical Review, 18(1), 65-78.
Humfrey, J. (1702). The Free State of the People of England Maintained. London.
Lucas, C. (1749). An Eighteenth Address to the Free-citizens and Free-holders of the City of Dublin. Dublin.
Miller, D. P. (1998). The ‘Hardwicke circle’: The Whig supremacy and its demise in the 18th-century Royal Society. Notes and records of the Royal Society of London, 52(1), 73-91.
Morris, C. (1741). A letter from a by-stander to a Member of Parliament. London.
Perceval, J. (1743). Faction Detected, by the Evidence of Facts. Dublin.
“The Humble Address of the Right Ho…”. (1705). The Humble Address of the Right Honourable the Lords Spiritual and Temporal In Parliament Assembled. Charles Bill and the Executrix of Thomas Newcomb. London.
“The Humble Address of the Right Ho…”. (1709). The Humble Address of the Right Honourable the Lords Spiritual and Temporal In Parliament Assembled. Charles Bill and the Executrix of Thomas Newcomb. London.
“The loyal catechism”. (1713). The Loyal Catechism; Teaching I. The Duty of Subjects to their Sovereigns. London.
Tolonen M., Mäkelä E., Lahti L. (2022). The Anatomy of Eighteenth Century Collections Online (ECCO), Eighteenth-Century Studies, Volume 56, Number 1, Johns Hopkins University Press.